This Week in Google 760 Transcript
Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.
0:00:00 - Mikah Sargent
Coming up on this weekend Google I, micah Sargent, am subbing in for Leo Laporte, and I am joined by Paris Martino and Jeff Jarvis. We start out by talking about how slog burnt slurf although slog didn't technically burn slurf, but anyway, you'll understand when you watch the episode. We talk about open AI and what we can expect from a new version of its GPT Plus Google getting hit with a fine in France for reading with its AI Systems. That, of course, sparks a heated conversation between Jeff Jarvis and Paris Martino about what it means if a machine Is doing the reading versus a human. Then we talk about Nvidia's recent conference, what we should care about there, what we shouldn't and, if you can believe it, we spent a lot of time talking about Dungeons and Dragons. It's all of that, plus so much more, coming up on this week in Google.
This episode is brought to you by Zscaler, the leader in cloud security. Cyber attackers are using AI and creative ways to compromise users and breach Organizations. In a security landscape where you must fight AI with AI, the best AI protection comes from having the best data. Zscaler has extended its zero trust architecture with powerful AI engines that are trained and tuned by 500 trillion daily signals. Learn more about Zscaler zero trust plus AI to prevent ransomware and AI attacks. Experience your world secured. Visit Zscaler comm slash zero trust AI. Podcasts you love from people you trust.
0:01:42 - Paris Martineau
This is Tweet.
0:01:48 - Mikah Sargent
This is this week in Google, with Paris Martino, jeff Jarvis and this week, micah Sargent. Episode 760, recorded Wednesday, march 20th 2024 the ground state of a molecule. This episode of this week in Google is brought to you by Rocket Money. Did you know? Nearly 75% of people have subscriptions They've forgotten about? Well, that was the case for me when I started using Rocket Money. I couldn't believe how many subscriptions I was paying for each month between streaming services, fitness apps and delivery service. It's never ending.
Thanks to Rocket Money, I'm no longer wasting money on the ones I forgot about. It was really nice to have this little bit of information that says look, you are paying way too much because, turns out, you have two of the same streaming service. What I what? How did I do that? I was able to cancel one of those and save some money. Rocket Money is a personal finance app that finds and cancels your unwanted subscriptions, monitors your spending and helps lower your bills so that you can grow your savings. I can see all of my subscriptions in one place and if I see something I don't want, rocket Money can actually help me cancel it with a few taps. They'll deal with the customer service for you. Rocket Money has more than 5 million users and has saved a total of $500 million in cancelled subscriptions, saving members up to $740 a year when using all of the apps features. Stop wasting money on things you don't use. Cancel your unwanted subscriptions by going to rocketmoneycomtwig. That's rocketmoneycomtwig. Don't forget rocketmoneycomtwig. Hello Paris.
0:03:53 - Paris Martineau
Love to pass the baton.
0:03:55 - Mikah Sargent
You did a fantastic job last week and we thank you for your work there. It was a great show and I am honored and very happy to be joining you this week. Also joining us is the Leonard Tau Professor for Journalistic Innovation at the Craig Newmark Graduate School. There it is. I'm Craig. I'm the school of journalism at the City University of New York and the director of the Tau Night's Center for Entrepreneurial Journalism at the Craig Newmark Graduate School of Journalism and the City University of New York. It's Jeff Jarvis. Hello.
0:04:30 - Jeff Jarvis
Jeff.
0:04:31 - Mikah Sargent
Hello there.
0:04:31 - Jeff Jarvis
I had to email Craig about something today and I said, oh, I've got to do my vocal lessons right now to warm up for singing your name, craig. I'm glad we've joined Paris. I'm very happy that you made this into a duet, thank you.
0:04:44 - Paris Martineau
Of course I have to participate. I can't leave you out there all by your lonesome.
0:04:47 - Jeff Jarvis
Yeah, we should try to figure out the I was going to say the lack of singing. We should do some harmonies.
0:04:51 - Mikah Sargent
Yeah, it really makes it work for me. It's sort of coming in at two different yeah completely both not synced with the Craig Newmark audio.
0:05:02 - Jeff Jarvis
We've got three completely different. Can we do it again?
0:05:05 - Paris Martineau
We're not going to be able to get it. We're on a zoom.
0:05:08 - Jeff Jarvis
We're rehearsed here. We're rehearsed One more time. Craig Craig Newmark.
0:05:18 - Mikah Sargent
Hey, the Newmark. The Newmark was almost perfect the Newmark choir. That's pretty good. Ah, this is the show where we talk about a lot of stuff and it is always a hootin' good time.
0:05:35 - Paris Martineau
That is a description for the show, yeah.
0:05:38 - Mikah Sargent
One of many descriptions I've heard for the show. At least, you know what I don't want to start with me this week. I would like for one of you to suggest the first topic. Is there something that is well? It looks like you might be saying Paris.
0:05:56 - Jeff Jarvis
But damn.
0:05:57 - Mikah Sargent
Paris is there something that you just really pumped to be talking about.
0:06:01 - Paris Martineau
Listen, all of my answers to this question are going to be the silliest thing in the world, so I'd like you to look within yourself to want to know if that's where you want to bring the show Immediately.
0:06:11 - Mikah Sargent
Let's go silly, why not?
0:06:13 - Paris Martineau
We've got a section at the bottom of the changelog called Paris' Brain is Hanging on by a Thread which you can peruse.
Let's see, there is something. I don't know if either of you guys subscribed to Matt Levene's column. It's fantastic. It's called Money Stuff, it's by Bloomberg and it usually has kind of a pithy, fond little headline that refers to one of the various news items he talks about this week. This week or, I guess, some perhaps yesterday. The title of his column was Slog is sorry, he burnt slurf. Do either of you guys understand that at all?
0:06:52 - Mikah Sargent
No, no no, is this a reference to something?
0:06:55 - Paris Martineau
It is. The headline has since changed, as you may notice while clicking on this, but no clear. It gets no clear. It has to do with basically a cryptocurrency issue, where you know a cryptocurrency called slurf, I believe Slurf may have. Let me find it actually here Slurf, a new Solana based meme coin. Slurf has faced significant challenges. After the project's developer accidentally burnt A major portion of the token supply, effectively losing 10 million, or the entirety of the presale participants' money Guys, I effed up the project's official ex-account wrote.
I burned the LP and the tokens that were set aside for the airdrop. I am so effing sorry. The slurf team later went on to a space, a Twitter space, to elaborate further on the situation. I'm sick to my stomach. Developer Slurfsoul said the issue comes in. Here is the Matt Levine title of Slurf is sorry. He burnt Slurf. It's because the original version of this newsletter attributed one of the quotes to a slurf employee, a developer named Slurf or Slorg. We asked for it, nika.
0:08:17 - Jeff Jarvis
You guys asked for it this is the world we're living in.
0:08:21 - Paris Martineau
So, after this newsletter went out, if you click on the tweet that I've linked in the rundown, there is a Twitter context pop up and it says Slorg is not the person who said any of the things that Mr Levine is claiming, nor is he on the slurf team. Slorg was a panelist on the spaces representing Solinsinator, and so all the comments are like Slorg saying bro, why are you using my name for this? I had nothing to do with this token other than the dude using this thing to burn the tokens on, and they had to issue a correction which is now at the bottom of that column which is perhaps my favorite correction of all time which says Corrects headline in 20th paragraph to clarify that Slorg did not burn Slurf. So that's where my head's been this week, woof.
0:09:13 - Mikah Sargent
This is always all just so stupid. I don't understand how anybody at all ever spends any amount of money on any of this. Is it genuinely just a hope that it's going to get anywhere near the value that what Bitcoin originally did? Why?
0:09:33 - Paris Martineau
do people keep doing this? People love the idea. People have been following for get rich quick schemes since the beginning of time.
0:09:41 - Mikah Sargent
Fair enough.
0:09:42 - Paris Martineau
Everyone believes that they're a special little boy or girl and that their time in the sun is coming. And that if only they just had the right combination of ideas or magic words to say that they'd be able to get rich too. And I feel like cryptocurrency has been particularly adept at swindling people along these lines. They think, oh, if you put all of your money into Slurf, you'll be a millionaire.
0:10:08 - Mikah Sargent
Because I'm a special little Slorg or Slurf, so I should obviously do it.
0:10:13 - Jeff Jarvis
So Sherrod Brown's opponent now is the Trump endorsed candidate in Ohio, who's a former car salesman, who's now a Bitcoin entrepreneur. That's all you need to know. There's three facts about this person.
0:10:32 - Mikah Sargent
I thought it was only NFTs that were this bad in terms of the stupid names.
I didn't realize that crypto had also gotten into the stupid name category. Is it not enough? I'm sorry, if someone came to my door 30 years ago and they were carrying behind them a bunch of Tupperware containers or Avan products, maybe I should give it a fake name so Avan doesn't get mad at me Avas products. And they said hi, my name is Slorg from Avas and I want you to buy this package of Avas products and sell it to people. But also, while you're selling it to people, you might also want to tell your friends that they, too, should sell Avas products. I'm not going to tell Slorg yes, because their name is Slorg. It's all these people that aren't real people and they don't have actual faces online, and you're buying things from a sloth named Slorg. Why would you do that?
0:11:33 - Paris Martineau
But, michael, you don't understand, is the dog with hat currency might be able to purchase an ad on the Las Vegas here. So that's huge. I don't the sadness in your eyes over just that second.
0:11:50 - Jeff Jarvis
The world is so broken.
0:11:52 - Paris Martineau
The world is so broken.
0:11:54 - Jeff Jarvis
What's even worse is the same people who are doing this stupid stuff are now in charge of AI, and they think they can destroy the world.
0:12:00 - Mikah Sargent
Yeah, yeah, because we're right, Slorg is probably. I should leave Slorg alone, because Slorg is unrelated to this. This is the Slurf community, the Slurps the Slurphazoids are probably the same people out there who are talking about prompt engineering and selling you their PDF that you can only buy with dog, with hat, coins, dog with hat.
0:12:27 - Paris Martineau
That's very important.
0:12:29 - Mikah Sargent
It's not with oh, it's like dog with hat, like a dog wearing a hat, not a dog sniffing a hat, or a hat that helps to provide sniffing for the dog, because I was picturing more like a hat. That helps you with things.
0:12:41 - Jeff Jarvis
Anyway, You're over thinking this back up. You are over thinking this.
0:12:46 - Mikah Sargent
Quick, joe, generate a dog hat that lets you with it. Anyway, my point is they're the same ones with the AI prompt engineering, which is a great segue away from this topic. That's breaking my heart into something about how AI prompt engineering is dead. Long live AI prompt engineering from you, jeff Jarvis. I didn't read the whole thing.
0:13:08 - Jeff Jarvis
Oh darn it it's long.
I put it in there, you caught me, line 55, thank you, where the hell are you? So everybody thinks the prompt engineering is the skill du jour, that everybody's going to do it, and so this story says that, no, this is kind of stupid. Absurd On the other hand, yes, we need to do it. The point is, as I read the beginning of it, that's all I read Is that we thought is that, basically, the machine can do a better job of prompting than we can. So the job we thought we had to prompt isn't the job we thought we had.
0:13:43 - Mikah Sargent
Yeah, I can like hands down. I can anecdotally, I guess, say this is true, because here's how I've seen this play out If I go to OpenAI's website and I go to Dolly and I type in a hat that helps you take whiffs of sense, it's going to generate something that looks very, very stupid and bad. And if I instead go to ChatGPT and I type in, give me an image of that very same thing, what ChatGPT does behind the scenes is it properly engineers a prompt, purpose built to get the best outcome out of Dolly Right, and then I get an image that looks fan, stinkin' tastic.
0:14:28 - Jeff Jarvis
So the nutcraft in this piece is that new research suggests that prompt engineering is best done by the model itself, not by a human engineer. This is cast out on prompt engineering's future, which is only like what? Three months old this job title and increased suspicions that are fair, a portion of prompt engineering jobs may be a passing fad, at least as the field is currently imagined. Nonetheless, there is a growing skill set in learning how to ask the machine what you want. However, if the machine, I think your other point, micah, is that the smarter the machine gets at figuring out what you want, the less smart you have to be in asking what you want. Yes, and I think that it will try, for you, but still there is.
0:15:08 - Mikah Sargent
I think that it's maybe not as because it used to be, for example, that if you were using what is it called mid-journey on Discord, that was a time where you needed to type in 15,000 different words to get exactly what you were looking for, to make it look exactly as you needed to. But yes, we've quickly realized that you don't need to necessarily like. That's what I want from this, and I think that that's what they're trying to create from this is that I don't need to figure out how to speak in its language.
That instead it should be doing what? Because I have prompt engineering for Siri that I train people on right. I have prompt engineering for Amazon's ALEXA that I train people on. So why is it not giving that result? Well, you have to say it like this no, we shouldn't have to do that. We should be able to say it however we want to. It should understand what dog whiff hat is whenever we're talking about it, which I didn't understand, but that's what it's there for.
0:16:09 - Jeff Jarvis
So we have a new job title it's whiff hat.
0:16:15 - Paris Martineau
It's not whiff, it's very important Together whiff Right.
0:16:22 - Mikah Sargent
Yeah, it's like whiff, but you can't say th, so you say whiff, whiff yeah, it's like a child's name Because it's a dog. Yeah, and it's wearing a hat, so it's a dog with a hat.
0:16:33 - Jeff Jarvis
You are overthinking it again, sir. Overthinking it again. Right? So there's a new job title which I had not heard before. Rather than prompt engineer, it's Lmups engineer, llmo, LLMOps engineer, which is what we're calling prompt engineers today. So you just give it a new title and you have to like your job.
0:16:55 - Mikah Sargent
Large language model ops. What is the ops?
0:16:59 - Jeff Jarvis
It includes prompt engineering, but it's more about Whereas it's in the life site. It also entails other tasks needed to deploy the product. So it's product development with LLMs. I mean Leo. So Leo of nine months ago I think it's about that time, paris, right, I was trying to find the original turning point episode and was having a hard time of it.
0:17:26 - Paris Martineau
So if anybody remembers that one, these guys have databases.
0:17:28 - Jeff Jarvis
They can find it. They're like Howard Stern they can find it all. But there was a time when Leo would say there's a passing thread, this is going to be gone, that's that. And then, at the turning point, he suddenly said no, it's the future I have walked.
0:17:42 - Speaker 6
Yes, I've taken the steps on the sand and when I or maybe it's Tuleman he walked with I don't know.
0:17:46 - Paris Martineau
Maybe, Maybe that's who he's on vacation with in Cabo. Maybe Lisa's just a front.
0:17:54 - Mikah Sargent
Wait, maybe this joke has already been made. Was there a footsteps in the sand joke already? No, oh, okay, because when there was only one pair of footsteps, the AI was carrying him. Is what I was imagining, anyway.
0:18:10 - Jeff Jarvis
What was? We invented a new metric when we did the NVIDIA right. How many?
0:18:14 - Mikah Sargent
Leos how many Leos are in a thing? How many Leos something? So they were talking. Take a step back here, yeah they were talking about these towers that are in these server rooms that have all these cool new NVIDIA GPUs inside, and he decided to use a carbon fiber vehicle. I'm bad with cars, so I don't remember which vehicle it was, but a carbon fiber vehicle as the metric for weight for this item, which means absolutely nothing to me at all.
0:18:52 - Jeff Jarvis
He admitted it, but he didn't even know how much it weighs.
0:18:55 - Mikah Sargent
Yeah, and so I looked at my favorite. One of my favorite things to do is go to and of course now I'm forgetting the name Wolff from Alpha and just type in any weight at any time, and what Wolff from Alpha will do is give you a comparison to something in real life. So if you and it doesn't have to just be weight it could be a length. So if you say 30 feet, that's equal to how many giraffes? So 120 pounds is equal to about three quarters of the average adult American female, or 0.88 the global average mass of an adult human. And we found out it was about I think we determined it was one and a half dairy cows was how much this tower weighs.
0:19:40 - Paris Martineau
That's a measurement I need.
0:19:42 - Mikah Sargent
And then Jeff said I wonder how many Leos that is, and so I used the average adult American male metric as Leos, and therefore we could sort of sub-Leo in, so 0.88 Okay how tall is Leo, though, and how tall is the average American man? I think we are going way off topic.
0:20:07 - Paris Martineau
Wow, you are trying to reign in the show.
0:20:10 - Jeff Jarvis
No, jeez, jeez.
0:20:12 - Paris Martineau
Last week we talked about Kara Swisher for an hour and a half.
0:20:15 - Mikah Sargent
I'm being told 0.58 is the average height of an American male. Anyway, let's move on. So now you are not an AI prompt engineer, you are a, a lemus. Why don't we have to give it a stupid name? And what does it mean? Does it mean that you understand how these interact with one another and so you are better at jumping between them, or something?
0:20:41 - Jeff Jarvis
Because I can understand that it goes back to the inflection discussion too. I think there is a what we used to call it my day was systems integrator. Yeah right, you take. Oh, we will buy those IBM machines and figure out how to make them work with the software we have in our operation. And so now these things are productized enough that you are taking something and you have to adapt it. The raw model will not do what you want. You have to do things with it, you have to give it data, you have to do other things. I guess that is what it is. So maybe it is fancy, nancy prompt engineering, I don't know.
0:21:13 - Mikah Sargent
Okay. Well, I think any number of people who both watch the show and are on the show could very easily be lamops is what I am saying, I think so too. But I mean that is the whole point right. It is all scammy, whammy.
0:21:30 - Jeff Jarvis
Well, in journalism, in the media industry, we have gone crazy, trying to pattern ourselves after Google, where we heard product is everything. If you are not product, you are nothing. So now the hot job in journalism is product Product development and audience development. Oh yeah, in organizations. Audience development. Also, product where you are going to make the Times Food app. You are going to work with marketing and technology and everybody else, but you are kind of the hub you are going to discover the next connections in the New York Times game out.
So my former school from which I am the orbiting, we helped start the international association of product people, and so it is a very big thing, very big.
0:22:17 - Paris Martineau
Yep.
0:22:21 - Jeff Jarvis
Alright let's move to product alliance.
0:22:25 - Mikah Sargent
Nupa, Nupa, nupa, dupe tiju. Where do we want to go? Next? Let's talk about.
0:22:35 - Paris Martineau
I thought it was kind of interesting that this week there was news about prosecutors are seeking kind of a 50 year sentence for Sam Bankman-Fried and his legal team came out swinging I'm on also line 89. His legal team came out swinging saying that that was a sentence only suitable for quote a super villain. Sam Bankman-Fried has said US prosecutors proposal to put him in prison for as long as 50 years. Quotes distorts reality and paints him as a quote depraved super villain.
0:23:10 - Mikah Sargent
I don't, that's 50 years, hmm, okay.
0:23:14 - Jeff Jarvis
What do you think? What do you think? What should he get?
0:23:16 - Mikah Sargent
Yeah, exactly that's what I'd like to know. What is appropriate for Sam Bankman-Fried? The non super villain, the villain, maybe that's what it is. It's the difference between being a villain or a super villain.
0:23:28 - Paris Martineau
I do think it's kind of interesting that it's far less than the 100 years recommended US criminal sentencing guidelines, but obviously much more than what Sam Bankman-Fried's lawyers say he should get, which is six and a half years. I believe it's partially because there were. The case involves over more than a million victims and losses of more than 10 billion dollars.
0:23:55 - Jeff Jarvis
That's a super villain. Apparently they're going to get it back.
0:24:00 - Mikah Sargent
Yeah, I'm not on the jury. I can say it. That's a super villain. I what? That? 10 billion dollars is a lot of money. Is what? Yeah?
0:24:09 - Jeff Jarvis
but evidently they're going to get it back because of his investments, so they won't lose money in the end.
0:24:14 - Mikah Sargent
But usually that's pennies on the dollar. Right, they're not going to get all of their money back, are they?
0:24:20 - Jeff Jarvis
I think that there's some speculation that they could Well that changes things-ish, but intent is still there, right? Well, does it? Yeah, that's the thing. Intent is still there, the deceit is still there, the lack of accountability is still there and, as Joe points out, you do need to create a deterrent effect.
0:24:43 - Mikah Sargent
Somebody needs to be made an example of right so that other people-. I don't think that's going to work with the crypto boys they're going to think they're immune. That's true. I mean, that's part of the profile.
0:24:55 - Paris Martineau
Interesting about this is I've got a couple links in line 87 and 88, which is as part of the sentencing Prosecutors posted at Google Doc, that they say that Sam Bankman-Freed wrote in the days just after FTX collapsed In which he mused about quote bad ideas for improving his image, and I'll just read some of these for you because they are-.
0:25:20 - Jeff Jarvis
These are bad ideas by you. These are bad.
0:25:22 - Paris Martineau
The Google Doc says in bold letters up at the top note these are all probably bad ideas that aren't vetted. Semi-colon confidential. Confidential.
You know there's some normal ones like file the doc with problems about the chapter 11 process. Go with the quote. The lawyers were in a bad situation and I feel really bad about that, like most people do. They reacted poorly to difficult circumstances. Number three go on, tucker Carlson. Come out as a Republican. Come out against the woke agenda. Talk about how the cartel of lawyers is destroying value and throwing entrepreneurs under the bus in order to cover up the incompetence of lawyers. We keep going on. Come out as extremely pro-crypto, pro-freedom. Go strong with the message of I have funding ready to make customers whole. If only the chapter 11 team would care about customers. Have Michael Lewis interview me on ABC. Go head to head with Matt Levine on odd lots.
Really lean into arguments All bad ideas, then if you go, farther down the Google Doc, he has a couple of suggestions of subgroups that he could reach out to support, including alt-right question mark and number two some other displaced group question mark God jeez. He has only one ally listed, which is Screlly misspelled, but I assume he means Martin Screlly. And under the random narratives part. He says quote let's see how he pleads as one of the narratives and the last one is SPF died for our sins.
0:27:00 - Mikah Sargent
So he's Jesus.
0:27:03 - Paris Martineau
You can't, you can't give Jesus 50 years in this slammer.
0:27:09 - Mikah Sargent
You can't call Jesus a supervillain. Let me out of here. I deserve six years at most. Ah, boy, I, yeah, I think again. You're right. Alright, so you're Judge Kaplan you're. Judge Kaplan knowing everything.
0:27:25 - Jeff Jarvis
You know, both of you. How many years do you give Sam McRefreet?
0:27:31 - Mikah Sargent
I don't want to go first.
0:27:35 - Paris Martineau
Paris, I don't know, I don't think I can make this a hypothetical and remove on one hand personally me Paris, the person, not me Paris, the journalist who has no professional opinions on this whatsoever Of course, of course more cypher Me, paris, the person.
I don't entirely believe in our criminal justice system in a penitentiary model, but yet also I think that white-collar crime should be probably prosecuted at a level more equivalent to how we prosecute other crimes. If we're going to, you know, take that into consideration. So I do think that there are situations like this where people need to face real consequences for actions that have harmed a lot of people.
0:28:18 - Jeff Jarvis
So what is sufficient to say? I better not do this because I could end up in years in the sing-sing.
0:28:29 - Mikah Sargent
I don't know. I think it depends on the person right.
0:28:32 - Jeff Jarvis
It, does, it, does. I mean one week would be enough to make me say, okay, okay, we'll do it, we'll do it.
0:28:37 - Mikah Sargent
I don't know anything truly about this guy, to know what that I guess the only I would say here's my ruling. I would let the judges, I would let the lawyers come forward, his lawyers, come forward and say this is what we think that Sam should get, and then I would add five to it. And that's what I would do, because the I think that's about right right.
0:29:05 - Jeff Jarvis
Six and a half plus five. Eleven and a half. I think that sounds great.
0:29:07 - Mikah Sargent
The number of years that they're arguing for is the idea that this person should get the bare minimum, should be treated, you know, as oh, he didn't do anything bad, whatever you did catch him. Oh, let's give him a slap on the wrist. So I'll take that number and I'll add something to it, and that feels fair to me.
0:29:28 - Jeff Jarvis
Yeah, if my lawyers put me in jail for six and a half years, I'd think. Can't you ask for less? Yeah, that's true, Six and a half, I don't know.
0:29:34 - Mikah Sargent
Hey, if we're going six and a half, could we not just go? I don't know, two years.
0:29:36 - Paris Martineau
You could make it five. Yeah, yeah, I mean, I think the thing is they have already a jury has already convicted him of multiple crimes. Yeah, they convicted him of seven charges, including wire fraud and conspiracy.
0:29:51 - Mikah Sargent
Yeah, those are bad things.
0:29:53 - Paris Martineau
Yeah, they are, it's not typically what you want to do.
0:29:55 - Jeff Jarvis
Yeah, no in Paris is right. White collar crime has to be more than marijuana.
0:30:00 - Paris Martineau
I do think that I mean marijuana. Shouldn't even at this point, but that's a topic the show doesn't need to go into.
0:30:07 - Mikah Sargent
I do think that Do you all not say that every day? Why?
0:30:09 - Jeff Jarvis
not, we go into everything else.
0:30:15 - Paris Martineau
I don't know. I think that the it is notable how flagrant he acted.
0:30:23 - Mikah Sargent
Yes.
0:30:24 - Paris Martineau
When it came to breaking the law, the existence, the reason I brought up this Google Doc 1 is because it's funny, but 2, it's like. What sort of person is like? Oh, my company has collapsed. I believe this was written after the bankruptcy proceedings that already started happening. I've lost like all of this money for all these people due to my lies. Let me sit here in brainstorm displaced groups and you know strange bedfellows that I can manipulate to get me back on top. That's a really disturbed way of thinking.
0:30:59 - Mikah Sargent
I agree with that 100%. And in that way I do think I mean, because what's to stop? You know you do this thing right and then you realize that you can do this thing and really not be punished for it all that much. Why not do it again? Yeah, I think some level of serious reckoning needs to take place. But, as you said, jeff, at the same time it doesn't necessarily mean that it's going to stop others from doing it, because part of that profile, so to speak, is to feel invincible. I don't know.
0:31:40 - Jeff Jarvis
Yeah, it's what Paris said. I mean, we're talking about two things at once SEMbeckman, Frieden, our entire so-called justice system.
0:31:47 - Mikah Sargent
So let's talk about something that made me think of you, jeff Jarvis, and that is Google getting hit with another fine in France. This time it has to do with France saying hey, we already have this whole issue where you Google profit, as France suggests, off of sharing news snippets from news publications when someone does a search. We also have a problem with you using the information from news organizations, from news publishers, I should say, to train your Gemini-barred model, your AI model, and so France is saying that requires that you pay us or pay, I should say, a 250 million euro. Fine, and yeah, again, it made me think of you, because this is sort of another segment of the earlier conversations about copyright, first and foremost, what they did in large part in Australia and kind of how all of that is shaping out in general.
0:33:11 - Jeff Jarvis
Exactly what so happens. I have a big paper, a 40-page paper, coming out about the California Journalism Preservation Act next month, commissioned by the California Chamber of Commerce. Full disclosure, and in there I talk about the prior tranche of fines. So, google, when the variant on the Leicester and Schultz-Rect the ancillary copyright the European version of this went through, france was the first country to put it into law. And Google said you got to let us, you got to give us permission to link to you or we're not going to. That caused a suit, and so Google agreed to pay $76 million to 121 publications over three years. Then the competition authority fined Google 500 million euros for negotiating in bad faith, which is to say negotiating, and this is around copyright. This is around Google itself linking to things. So now we go to, and so the entire news industry this is what my paper is about the entire news industry is trying to shrink fair use and expand copyright. They're trying to say that linking to us and quoting our headline is not fair use. Okay, google, stop.
0:34:27 - Paris Martineau
Google is trying to fight back against you.
0:34:30 - Jeff Jarvis
They are. They are yeah, it's on mute. Shouldn't do that, Sorry. So they're trying to expand copyright and shrink fair use when it comes to linking to their stories and giving them value. They're not acknowledging the value of the links.
Okay, so now we advance forward to what I testified about in the Senate, which is using content for training models. Now here's where it gets interesting. Well, it's already interesting, but here's where it gets devious. Is that what, what the New York Times is saying? What they're saying in essence is Never mind going behind paywalls. If you read our content and use our content, Then you must pay us. Well, that's an entirely new doctrine of copyright that says to the machine and the machine's owners you can't even read something that's copyrighted, because everything in the world is copyrighted. Everything you write is copyrighted automatically. Now, so they're trying to argue that if you just read it and use it, you owe us money. And so the French government is coming down on the side of the publishers in this. What I would say is Foucaque the argument, and so finding Google for $270 million, and I think it's completely wrong.
This is what I said before the Senate, and it is the news industry that is ruining freedom of expression. It's ruining the right to learn. So one little plug for me I'm holding an event in New York on April 30 with the Common Crawl Foundation about just this topic. It's right to learn. And AI and the open Internet. You got an invitation, Paris.
0:36:09 - Paris Martineau
I'll be there, yay.
0:36:12 - Jeff Jarvis
And if you want one, you can let me know and I'll see what we have in terms of space. But this is a bigger issue. So the French government is siding with old publishers saying oh yes, yes, your stuff is so precious that that it shouldn't even be read without paying you. That's BS and I hope this goes to the court of human rights and I hope somebody stands up for freedom of expression and the right to learn and the right to read. Thank you for that moment of ranting.
0:36:38 - Mikah Sargent
You're very welcome. I the idea that it the idea that there's a separation right between me going and reading something versus this thing going and reading, reading something, taking it in on its own, that starts to get so messy, so quick and and I can't remember was it us who? I was talking to somebody about micro transactions for sites, but the idea that, as you're going around looking at a site, that every little thing that you do on the site, every link that you click I must have been Steve Gibson would cost you money and in a world where you start to go down this path of you know, discounting the right to read, that I or anyone doing anything anywhere online would be not protected as as an action is. I don't know. That's. That's just a complete breakdown in my mind of of everything that I don't know I certainly hold dear.
0:37:51 - Paris Martineau
I'll push back a little bit, because I agree with your sentiment, micah, but I don't and this is a point I feel like I've made in the show time and time. I mean, I don't think me as a writer is fully in this. I think it's just a misinterpreted argument in some sense. I don't think that a large language model, a, a model created by a commercial entity that exists for commercial purposes, has any rights to anything, much less to to read, which in this case means Consume content and train itself on it. I do think that there are, you know, more complicated questions to be asked here. To say, do for-profit entities have the right to use other people's copyrighted material to Enrich themselves?
0:38:44 - Jeff Jarvis
so, paris, you work for a commercial entity called the information, which charges a lot of money to people, including me, and you read the New York Times. You have a subscription, I imagine, and you believe that's part of your right. By reading the New York Times, you can use information. You got there, not direct phrasing, but you can use that and you can read people's blogs and you can talk to sources and you don't pay any of them for that and you don't have a specific license to use that. But that is the essence of fair use is that you have the right to learn from that, because information is, is Is free.
0:39:20 - Paris Martineau
That sense information that a computer program Necessarily has the same rights as me a person well, look out this way.
0:39:31 - Jeff Jarvis
So how many? I had this somewhere. I think in the rundown a lot of the Pulitzer Prize nominees this year Are did I keep that in there? Five of this year's Pulitzer finalists are AI powered.
0:39:44 - Paris Martineau
What is so and you know why that is? It's because most often the Pulitzer Prize finalists or winners are from like the same three incredibly large newsrooms and the People that end up winning those awards are on teams when their whole job is to like work on a couple of stories tailor-made for awards over the course of a year.
0:40:04 - Jeff Jarvis
So, of course, they're going to use every tool at their disposal right, which in certain cases includes AI right, and so I would bet one of these is probably the New York Times, and the New York Times is suing open AI for doing what the New York Times does, which is to gather data and analyze it in the machine and then use it for their purposes, which are commercial. And that's what I'm trying to argue is that, is that Forget about the machine versus the peer person, the company versus the company. New York Times is a commercial enterprise. The information is commercial enterprise. Open AI is commercial enterprise. That's makes an even playing field. They're all commercial enterprises.
0:40:35 - Paris Martineau
But I think part of the difference is that, in the case of something like open AI, what is being sold is explicitly the machine that is trained on the use of this machine that is getting smarter because it is trained on a Specific data set that is growing ever larger. In the case of, say, like the New York Times or some other a journalistic example, an AI model is being used to go through a large data set which might be One sentence, a reference to which will be like one sentence in a story that is like of three to four thousand words. I think it's very different because ultimately, in the case of like the New York Times, the commercial aspect of the enterprise is you're paying for a subscription to a body of work produced by thousands of journalists. Versus Open AI, you're paying to use a commercial tool that derives its value from the work of others.
0:41:33 - Jeff Jarvis
Drives its value from isn't the information. That information, by the way, is free, because you can't copyright information in an enlightened society. What it's driving the value from is grammar, it's the, it's the Coincident placement of words and the tokenization that occurs as a result.
0:41:50 - Paris Martineau
So it's learning how to speak, so that skill is speaking grammar is all that matters, then why doesn't you know, open AI and these other places just totally Swerve this issue and exclusively use uncut, like works where copyright is expired to train all their data set? It's because? The answer that is because they want up to date information so they can provide people with Correct, relevant new info, because that's what they're getting out of it. Well, they search for something.
0:42:19 - Jeff Jarvis
But of course, the models are stale and you don't actually get current information. This is where I've been arguing from the. What I've interviewed the CTO of shipstand, the Swedish publisher or Norwegian publisher, both on AI inside show with Jason is I've been arguing that we and new should create an API for news, that we should make our content available to them with a deal and say here, here, you have a key to this API to get current information when you need it. And that's what's happening in Europe. I really wanted to go. There's a. There's a conference in April on on AI and media In Copenhagen next month with it they can't go to, which is killing me because they have a different attitude.
They're trying to say when can we mutually benefit here and use this going forward? And Norway is building its own LLMs with its own stuff to do that. It's a smart way, whereas what's happening here is the New York Times is suing them, say get out, stay out. Well, the truth is New York Times is fungible. Open AI doesn't have to use the New York Times. It can use lots of content, lots and lots of it can use business insider Right when they've done a deal with with it is yeah.
Right, they've done a deal with with Springer. Axel Springer yeah, and what does business insider do? It reads the New York Times of the Wall Street Journal. We writes them, so would get the same information. It would get the same associations. Otherwise they could do that. Yeah, they're probably opening eyes.
0:43:45 - Paris Martineau
About to get really into five bullet point summary.
0:43:51 - Jeff Jarvis
And and pictures of. I took a train trip and let me give you 40 photos about it.
0:43:56 - Paris Martineau
I will say I Was friends, a lot of people who worked a business insider and, unfortunately, those articles that we all love to hate, those make them so much money.
0:44:05 - Jeff Jarvis
Oh yeah, I'll click on them to hate exactly because it's the easiest way to get the information.
0:44:09 - Paris Martineau
I went to a random town in.
0:44:11 - Jeff Jarvis
Wisconsin.
0:44:12 - Paris Martineau
Here's what I thought about it.
0:44:15 - Jeff Jarvis
And it somehow sucks us in. But my kid is another point too, I think, which is one of the ironies or paradoxes or something like that of this, is that the NFT boys and the crypto boys and the blockchain boys they talk a lot about ownership Everything should be owned, which is kind of your conversation with Steve you come across anything, I own it, you got to pay me for it. Everything is a bridge with a troll and a toll and you got to pay for it. And that's the exact opposite of the ethos to me of the internet is information is out there and we can all take advantage of that information, and that you can't own information itself. So there's a fight going on here, not just among the old publishers, but even among the new boys here around this, this notion of ownership.
0:45:01 - Mikah Sargent
Yeah, because it sounds like. Really what you're saying is there's a there's a separation between Information, which you perhaps are suggesting should just be. You know, as you're saying, it's, it's free, it's available to everybody. How Does that reconcile with, especially for for the career I used to have, paris's current career, like how? How does journalism Exist in a world now I'm playing like both sides here how does journalism exist in a world where Information is fully and truly free online? Is that possible?
0:45:46 - Jeff Jarvis
So I write about this in one of my earlier books where it's the diamond and water paradox. Diamonds are pretty much useless but they're incredibly valuable. Water is absolutely necessary, but it's almost free, mm-hmm. And it comes around. Similarly, around information.
And I happen to have in front of me the famous quote from Judge Louis Brandeis about just this issue, dissenting in INS versus AP, that the general rule of law is that the noblest of human production knowledge, trues, ascertained Conceptions and ideas become, after voluntary communication to others, free as the air to common use. Now, a part in phrase there is after voluntary communication. No, there's not. After theft. Theft would be I go behind the paywall and I take that. I take your publication. I don't pay you for it. That's, that's a fair issue to have.
But once the information is out there, once I've read it and I repeat it to you on a show like this, right, I read the information before we get on the air. Yeah, and it cost me. I got a good, good discount of $299, but I repeated all the valuable information in that article and that's my right. That's fair use. That's perfectly fine and legal. But we won't let a company with a machine do the same thing and we're all gonna use that machine and but it's alright, paris, we tried again. You say the same thing about search engines and social media. To look at the at the California Journalism Preservation Act, they are saying that you shouldn't even link to and quote a headline of a news publication In either search engines or social media unless you license and Pay for that rate. What do you think about that? I?
0:47:28 - Paris Martineau
mean I disagree with that because I think that it's Ridiculous. I think we can all agree that a headline and a link which would take you to somewhere's one site where you can choose to pay or not pay is it's fair game and it's mutually beneficial.
0:47:46 - Jeff Jarvis
So again said to me, what's different from the deep mind division of Google versus the search division of Google? I?
0:47:52 - Paris Martineau
Think that one is consuming the entirety of other people's content and Basically creating its own value from that so Google the other is aggregating links.
0:48:11 - Jeff Jarvis
Google to the search engine to work, it aggregate. It comes in with all of that content, not just the headlines. It Absorbs, it crawls all of that content.
0:48:22 - Paris Martineau
It makes just the content that is available to the crawler. A lot of people don't have all their True that, true all around if you follow.
0:48:29 - Jeff Jarvis
Do not follow links. That's, that's fine, right and robots a text. But what Google does is in fact read the entire article so it can understand how to group news articles together, and understand that it may only display the headline, but it read the whole article, it read the entire site. It did exactly what you just said to create a good search or news experience.
0:48:54 - Paris Martineau
Michael, what do you think on this? You're being very Switzerland, jeff, and I talk about this. We have this same argument every single.
0:48:58 - Mikah Sargent
Every time. Yeah, yeah, ultimately, I think I'm more, I'm leaning towards the Jarvis camp, from, yes, from the perspective of, from the perspective of, I guess it's. It almost has a more for me, a more rose, tinted shades kind of Feel to it, and I tend to be more the optimist. But the interesting thing is and I say that optimistic in terms of Information being free and wanting it to be the case that people can Read a story and share that information and that that does not come, you know, but there's there's, no, there's no issue there, or, even more so, that we have Social media platforms where people can make stupid, stupid jokes with images from shows and that that doesn't have to be Something that the platform has to pay for and get in trouble for, which means that basically all of the means of social media that we have would get closed down in that as a defending Joe Esposito.
Let's be clear here right, exactly I.
That is some of the stuff that I value about the web, and it's some of the stuff that I see as the important aspects of the web, and the thing is that, underneath those rose tinted shades, there is also that realism aspect too, because it's happening and it has happened and it continues to happen, and, as I think, also as Jeff points out there at the end, the idea that Web crawlers are doing their web crawly thing, not to mention that that Robotstxt is not an official, agreed-upon, legal or even like sub legal standard.
That's just something that people handshook on years and years ago. I just jostled the microphone for the fifth time, and so web crawlers tend to follow that, but there's nothing that says that they have to, and so I wouldn't be surprised if there's even, you know, further outreach. That is, yeah, that's where I stand, but at the same time, I Do think that if something is behind a paywall, is locked off as unable to be seen, and that it you know, the agreement is, you pay for it to get to see it, mm-hmm, that stuff should not be Taken in.
I agree with that, and so I think that maybe in the New York Times.
0:51:34 - Paris Martineau
Jeff well behind the paywall.
0:51:36 - Jeff Jarvis
Let's put it, let's put this way no, the New York Times material that was on that they met demanded to be taken down from common crawl was out and public for free. And and well, I'll argue the other way. Then I subscribe to the New York. I'll I need is one subscription of the New York Times. To buy the New York Times. All I need is to buy as one book to read the book. Yeah, okay, fine, I'll go to that extreme all the. All the opening I has to do is to have one subscription of the Times and I should be able to believe I've smoked, talked you into being more radical as part of this this is how extremism happens.
Peace by peace. The Paris rabbit hole. We've gone down a Paris tunnel just now.
0:52:18 - Paris Martineau
Oh Wow, hey, this is, you know I saw.
0:52:20 - Benito Gonzalez
I've been listening to this argument for the last couple of months now.
0:52:27 - Paris Martineau
No, and like I mean.
0:52:28 - Benito Gonzalez
I've always fallen into sort of in Paris's camp, and this is mostly because not really because of the training and all this stuff, but because of I see the ultimate end goal of these companies is to ultimately cut us out, and that's sort of like the. The problem that I see in this is that this is what they want to do. They want to make it so that there are no more journalists, so that the computer is writing everything, and this is sort of like the. The issue that I have with it. It's not necessarily the training, data and all that other stuff. It's like what their end goals are. That's what I don't I totally agree.
0:52:57 - Jeff Jarvis
I, you, so we have camp against camp here. Like you have to come back every week now so that I have an ally.
0:53:05 - Paris Martineau
Leo Laporte.
0:53:07 - Jeff Jarvis
No, leo often disagrees with me for the sport of it. So Not necessarily a story for us, but a quick mention that both Gannett and McClatchy, to the three largest newspaper chains in America, just said yesterday they're gonna stop using the Associated Press, stop paying for it. That isn't just a matter of them stopping to pay for it, which is bad, because where are they gonna get their news from when they don't care, because the crappy papers owned by hedge funds.
0:53:33 - Paris Martineau
All a significant amount of their news. It's worth pointing out. Time has come from the Associated Press of exactly tons.
0:53:40 - Jeff Jarvis
Where are they? But the other part of this is the Associated Press is a cooperative, so members of the Associated Press give their content to the network for the good of all, and I presume they're dropping out of that as well. And so there's this clamping down where the news industry figured out how to share among itself. They used to fight at first. First they sent boats out to meet the boats from England to grab the news first, and that was expensive because they had competing boats. So they agreed, honest to God, to start ready the same boat to get the same stuff.
0:54:09 - Paris Martineau
That's basically the beginning of a wire service and along comes the boat, just to breathe a sigh. I think we did that. Yes, bring it back.
0:54:16 - Jeff Jarvis
Is there a river near near pedal away? We could we have a?
0:54:18 - Mikah Sargent
little look down on the news boat. Yeah we got a little river.
0:54:22 - Jeff Jarvis
Then came the telegraph, and so they needed to figure out how to do that, and they tried to keep out, like Hearst wanted to keep people that's are out of it. And then finally they agreed, okay, okay, we're gonna have a wire service together, and that's how the AP came, and now we see these big awful companies tearing it apart the same thing Murdoch did in Australia, where he pulled out of the wire service and thus it lost value. So we've got to think too about, especially as journalists, our obligation to the larger news and information ecosystem in society. In the paper I'm writing about the Legislation in California, one of the ideas up for right now I'm trying out a whole bunch of ideas. Well, one is to say, okay, all right.
If, if Google or tax dollars are gonna create a fund for news, why shouldn't that be used to buy content out of the paywall for the good of all the public? If the public's gonna pay them the freight, then why should the public benefit from it? Everything is going behind paywalls right now, and fine, you got to earn your living, that's okay. And the information is unique in this because it has unique information, unique reporting, but the average connect paper crap.
0:55:30 - Mikah Sargent
We do have to take a break. I'm realizing.
I do know as thankful. So yes, This episode of This Week in Google is brought to you by CacheFly.
For over 20 years, CacheFly has held a track record for high-performing, ultra-reliable content delivery - serving over 5,000 companies in over 80 countries. At TWiT.tv we've been using CacheFly for over a decade, and we love their lag-free video loading, hyper-fast downloads, and friction-free site interactions.
CacheFly: The only CDN built for throughput! Ultra-low latency Video Streaming delivers video to over a million concurrent users. Lightning Fast Gaming delivers downloads faster, with zero lag, glitches, or outages. Mobile Content Optimization offers automatic and simple image optimization so your site loads faster on any device. Flexible, month-to-month billing for as long as needed, and discounts for fixed terms. Design your contract when you switch to CacheFly.
Cachefly delivers rich-media content up to 158% faster than other major CDNs and allows you to shield your site content in their cloud, ensuring a 100% cache hit ratio.
And, with CacheFly's Elite Managed Packages, you'll get the VIP treatment. Your dedicated Account Manager will be with you from day one, ensuring a smooth implementation and reliable 24/7 support when you need it.
Learn how you can get your first month free at cachefly.com/twit. That's C-A-C-H-E-F-L-Y dot com slash twit.
All right back from the break, and I do want to say this is maybe we'll move on from this, but it is a hard thing when you it's a hard but very real thing when you do consider the human beings that are potentially impacted by some of the aspects of what's around this argument or discussion that we're having. Right, because you know, I have long, I think, rolled my eyes at the. There's a huge swath of, especially digital art creators who get a little up in arms about what generative AI is able to do. And then there's the conversation about how gen AI will never reach the level of not perplexity, but will never reach the level of complexity that it would even come close to a human. You'd always be able to tell between a human and something that's created by generative AI.
But increasingly we are seeing many a publication decide to pick up generative AI content, especially for some of the very quick stories that they can write about I don't know sports outcomes and other things that don't require much brain power. And, yeah, that's where I fall in with what you're saying, benito, this idea that those of us who have long held a job in a field that is perhaps being threatened by this. But I also think, jeff, that that's almost a separate argument from the general argument that you were making, right, I mean that maybe it's not an argument, it's more of a point of consideration, but it has nothing, or it? If we can think about things in a vacuum, which we as humans do not do, then there is that's where I really fall into that camp of understanding what you're saying which is that if we start saying that a machine can't read this thing that's available to everybody else, where do we draw the line? Does that soon mean that you know we're setting more rules? And that exists here.
But then you also look over here and you've got five of your friends who work in journalism I guess, sports journalism specifically who have these sad faces on, who are going AI took my job and then this starts to fall down. And I guess that's where it's very complex and where it is interesting hearing from the both of you and your perspectives, because that is also where I understand Paris your point about like, this is a company who makes money by using information from a company that needs to make money off of its own information, and so if it is going to that company and taking that information for itself, where you know, should that not be paid for? Or what have you? I get, I get all the perspectives. Is what I'm saying.
1:00:14 - Paris Martineau
I think something also that kind of goes and said often in our debates on this is that, which I think is a really good point, micah, is that part of what is motivating the reactions from, I guess, regulators on this, as well as kind of the outcry, sometimes to a ridiculous extent, like the outcry over like headlines or links that you mentioned, jeff, part of what is motivating this is a feeling of well, we we as an industry didn't really we are, I guess, a society also didn't fully catch the tech companies and regulate them when we should have.
There is this whole moment that we're in of kind of animus that is motivated by well, we screwed it up. You know, we let all of these companies grow so large, accumulate a large amount of social and political power, and now we are starting to see a new industry grow and the inclination is to crack down hard to make up for our past mistakes. And obviously I don't think any of us can say whether or not that is a good instinct that will bear out. Well, I'm sure we will probably just make as many mistakes in the future as we have in the past, but I think that it's worth, you know, interrogating a bit. It is the idea that people want to be a lot more proactive when it comes to regulating industries that could enact sweeping changes across the rest of industry broadly.
1:01:51 - Jeff Jarvis
It's a little rich for folks in media to complain about growing big and powerful on controlling power, when that's the essence of what media have done. And media tried it in the paper I'm writing Well.
1:02:08 - Paris Martineau
Jeff, I do the history. It's rich for you to say this, looking at the places where you have worked over your career. Yeah, I agree.
1:02:14 - Jeff Jarvis
I agree, I agree, I absolutely. What I'm saying is that I think that we have to be self-reflective at this point about what media have done in society. And that's what I'm really asking for here is that if you want a fair playing field, then if we read other stuff and get information, then social others people do that. If we get criticized for holding on to power, if they get criticized for holding on power, so should we get criticized for that. But, of course, who are the ones who do the criticizing Us in media? Who don't we criticize Us in media and the hegemony of media.
Going back to when radio arrived, newspapers tried everything they could, exactly like the script is exactly what it is today. They tried to expand copyright and shrink fair use. They tried to get the new medium regulated and they succeeded in cutting a slice out of the, out of the first amendment with the FCC. They tried to prohibit radio journalists from competing with them. They tried to prohibit the new medium from getting an advertising competing with them. They complained about how it was hurting them until enough of the radio. The newspaper owners bought radio stations and then it was all okay.
And so we've got a very checkered history in our industry and, yes, I'm part of that. I've been part of that industry my whole life, my very long life. And, yeah, I think that we should start by looking at home, and part of the problem here is that, micah, you're right to be certainly empathetic to the journalists who are affected, but the blame in my mind starts not with Google. It starts with their own incompetent greedy owners, who got up too much debt in their consolidations and then became vulnerable to hedge funds buying distressed debt, and that's why they are where they are right now. And so they're using their lobbying money to try to get money from Google and get meta.
If meta doesn't prop news, like they did in Canada, to pay the big old hedge fund owned companies, any news organization that doesn't already earn $100,000 a year is counted out. So all a whole mess of diasporic and black media and startups and community media won't get a penny from this stuff because the lobbyists don't care about them. The lobbyists for the hedge funds are going for the big guys. So, yeah, we can complain about the nastiness of big, big corporate life in the platforms, but let's look in a mirror too. As a whole, I like the information. I'm not saying the information, but I am saying as a whole the media industry is sucky. Yeah. I think we can all agree with that.
1:04:53 - Paris Martineau
I think, we all agree with that.
1:04:56 - Mikah Sargent
All right, then let's move on, because I do want to talk about NVIDIA, NVIDIA's conference, so NVIDIA add what I still feel is a very techy, very nerdy conference the GPU technology conference and Jeff Jarvis and I happened to cover the event here on Twitter, and at this event, NVIDIA announced some new hardware the Blackwell chip, as well as sort of the overall Blackwell platform, which is this whole system of chips and hardware that all come together in these huge devices that again weigh about two and a half Leo's and they are supposed to be more energy efficient.
But what was interesting is that they it wasn't like they're more energy efficient so you can feel good about not hurting the environment with all of the AI stuff you're generating. It was no, they're more energy efficient so you can use more of them at once to do even more than you've ever been, than you've done before, Plus a whole lot more. So I thought we could kind of break into it, first and foremost with the Blackwell portion of the program and how they're kind of combining this new Blackwell architecture alongside the Grace CPUs that they have, although they did ditch the Grace Hopper GPUs in favor of this newer technology Named after a mathematician, I believe it was.
1:06:39 - Jeff Jarvis
Grace Hopper is a famous pioneer of computers. Let me try this out on Paris. But Blackwell. He holds the Blackwell chip, named after a Black mathematician, yes, mathematician, david Blackwell. And then he holds the Grace Hopper chip and the one is going to supersede the other, and then he looks at the Grace Hopper chip and says, good girl, yeah, that's what we said too. Yeah, oh, same reaction. Oh yeah, yeah, that was, that was bad.
1:07:09 - Paris Martineau
I was going to say that anecdote is the only thing I know about that conference. To be totally honest, my question and.
I had heard that. Yes, and that was my earnest and actual response when I read it for the first time. As well as perform, for you guys live right now Very good work, but my question is this is perhaps a bit of a faux pas to me why should I care about this as someone who historically my brain, I feel like turns off when I hear about chip developments? Why should I care? Is it just that they'll be able to do more faster?
1:07:47 - Mikah Sargent
I love that question because I mean, yes, you're right, At the very base of it, it's that they'll be able to do more faster. But I think the only reason you should care and maybe it's something that you, once I tell you you're going to be like, well, I care even less is because, as we know, NVIDIA is at the heart of a lot of the generative AI stuff that's out there in the world right now. So any improvements to this technology likely means improvements to all of the Gen AI tech that we're seeing across the board. Amazon, Microsoft, Alphabet, Google are all different partners that work with Blackwell, and so we're bound to see lots of different services start using these in the background, given that they're already using the current run that has the hopper chips in them.
1:08:43 - Paris Martineau
So I know that a big point of contention. That is interesting. Yes, I know that a big point of contention for when it comes to these chips is availability or the lack thereof. How, what's going on, I guess, with the supply chain for these. Is it realistic to assume that these chips will be available and in use by all this company, as you just mentioned, imminently, or is it again going to be a supply chain issue like previous iterations?
1:09:10 - Mikah Sargent
That is a really good question.
1:09:12 - Paris Martineau
He didn't address that.
1:09:13 - Mikah Sargent
Yeah, he didn't address it on stage. I'm reading in this piece here that NVIDIA has since said that they are going to be available later this year to the companies that might be scooping this up. But this is like so far up the chain kind of purchases. These are very, very, very expensive products that it was clear, were, you know, it's layers and layers and layers of difference between what we would buy versus what they would buy at the line. So I imagine they don't need to make as many of them. I guess, is what I'm saying, but yeah, it's.
What was interesting, jeff too, is that they did speak on a very short scale. Everything that and remind me of the CEO's name, oh Wang, yeah, wang was saying was five years, it was all about five years and that over the course of the next five years, these are all the things that we're planning on doing at NVIDIA. This is what's going to be changing at NVIDIA, and brushing up against wanting to utter the three letters that many of us are kind of going okay, which, of course, is AGI. We felt that kind of happened a few times where there was hinting at that, but I also yeah, exactly, didn't say it, but hinted at it.
I also found it interesting how often we saw the Apple Vision Pro pop up in this keynote. It was there's a lot of talk about, I can see Paris is why Exactly you think what's going on. Well, there was a huge part of the show devoted to, and then sprinkled throughout afterward, a concept called a digital twin, which is not a it's not like a K-pop version of me. No, sadly, it's not a K-pop performer version of you and me. It is instead a sort of far more massive scale.
A company can create a digital version, for example, a manufacturing plant that they plan to build, and then, in the digital version, they can not only test out workflows and simulate things to see how, eventually, it will run and what would be the best opportunity for it, but the company ended up or the company NVIDIA ended up talking a little bit later, which we'll get to robotics and using these digital twins to train the robots in this virtual space before they're brought to the real world.
But earlier on, when they first started talking about digital twins, they were using Apple Vision Pro as a way to tour your digital twin workspace so you could walk through and see how the manufacturing plant was going to be built up and I made sort of the joke to Jeff that I felt like this was going to bring forth a whole new era of Karens and Kevin's I guess Sorry, kevin which the people who are wearing these Apple Vision Pros walking up to the contractors and saying, no, you can't put that sewer pipe there, because this is telling me that's a bad place for it, not having any idea what they're talking about and that seems like my nightmare.
1:12:36 - Jeff Jarvis
Anyway, Let me go to your question, paris. Why should you give a damn? I'm somewhat the same way. But the early, very early days of Twitch I would tune in and they would have a discussion about the specs of a new monitor and I would say, why should I give it to them? I really don't care Does it turn on or not. Right, but this is that point. They cared about the hardware and it was interesting and of course then it became very boring and we didn't care about that anymore.
1:13:02 - Paris Martineau
Somewhere out there, a lot of listeners are going.
1:13:05 - Jeff Jarvis
hey, I care, I care about the specs of my monitor, Derek Drevers, Don't you kill my nits. But so what happens, obviously, is that is that high technology and hardware view, I think, fades in the background. Clay Scherke famously said the social media gets interesting when it gets boring what people do with it. I think the same was true of other technologies. So I was the same way.
I didn't really envy the chips there. You know, they asked me to do the thing with Micah and I like Micah. So okay, I'll do this, I'll learn something about it. It's fine and a lot of it I didn't understand. We both admitted going up, we didn't understand a lot of it. What struck me with those was this Paris is that the scale at which they think is awe inspiring in good ways and bad, I mean fine, good that can think big. But my Lord, when they started saying what all this one chip could do, how much exponentially bigger and faster and more transistors and more pedophlops of floating point transactions, it could do. 30 billion trillion floating point transactions, 30 billion trillion flame, whatever the hell that means.
1:14:20 - Mikah Sargent
Right, I was gonna say just that is. How many dairy cows is that? I don't know.
1:14:25 - Jeff Jarvis
Yeah, and then you put two of them on this board that looks like a robot with two chips and then a nose and then a mouth, and that costs a fortune, right? And then you put a bunch of those boards in a device and then you put a bunch of those devices in a rack and then you put a bunch of those racks in a row and then you put a bunch of those rows in a data center and they're illustrating this growing and growing, and growing and growing. When you think about the scale of that one chip and now the multiplication of it across all of that, and then Huang says now you got to fill all that up and that's when we get back to our data conversation is that they've already sucked up all the possible human data they can get. Now they're going to make up a whole bunch of data just to have data, just to fill these things, just to do these things. And this brings me back, as it often does, to the Stochastic Parents paper where they screamed about enough of you boys with your size matters making the ever bigger models.
That's dangerous. It consumes electricity, it ruins human lives who have to work with it, it can't be audited for what goes in or what comes out, stop. And I felt a bit of that where I wanted to scream at some point. I didn't because I didn't want to blare into my case here. Stop Enough. Already it's bloody big enough and I understand that's stupid. I understand they're going to go ahead and build big things. They're going to build bigger things. That's the way the world operates. I understand that.
But the scale of it in some about a two or about a third of the way in the presentation, I was saying, oh hell, this is big and it's going to have implications.
Now, at the same time, I believe that open source, small models are going to matter a great deal because they're going to enable researchers and small companies and other countries and other languages to work, and that's where I want to see the development and that's what Yanlacun talks about more at Metta. And I think that the big mall I agree with the Stochastic Parents papers authors of the big models are crazy in a lot of ways, but oh my God, they're going to get bigger. So I think that there was a and I don't look at this in a policy sense, I just look at this in a societal sense of saying these things are going to get really huge. How are people going to imagine using this computing power that we can't even imagine now? And I'm not a moral panicking, I'm not suggesting it needs to be regulated, because I don't know what you're regulating, but I am saying we got to pay attention. Is that an answer?
1:16:55 - Paris Martineau
That is an answer. It was a beautiful monologue, thank you, I'm contemplating all of those rows and rows of chips, if you even looked at that.
1:17:04 - Jeff Jarvis
we should find that one like three minute part of the presentation, not just today. Just get it zooming out and zooming out.
1:17:09 - Mikah Sargent
Yeah, it was all chute and they said it's your job to fill all of this with data processing essentially. And they explained how one word is essentially three. What are they? Tokens, tokens, tokens, thank you. And so a phrase and they used a phrase from Star Trek was 30, 40 tokens, and that, actually, that was a moment for me where I thought, oh my goodness, sometimes I get really wordy when I'm talking to these GPTs and so I'm just imagining how many tokens I'm taking up at a time with some of the things that I ask of these GPTs. It's very interesting.
This was an opportunity for the company to flex, and for the company it wasn't even just a flex, it was like a peacocking flex. And I say that because, again, the part where I mentioned earlier the backbone of a lot of the processing and the creation of and the work on different AI models, but the consumers especially and then those of us who are sort of one level away from consumers, being that we have more knowledge than the average consumer about this stuff also may not realize it. We think about Microsoft and its co-pilot, we think about Google and its Gemini, we think about Open AI and its chat, gpt and other tools, and so I think that it was an opportunity for Nvidia to go. Don't forget about us. You all are getting all excited about this and we love that. Let me tell you about how that's us in a lot of the way, and then also, at the same time, show what we have planned for the coming years past just making better GPUs, better, bigger GPUs and they showed, as I mentioned, robotics technology.
Again, this had a lot to do with that digital twin. I was talking about where you would have a robotic factory, like at Amazon, and if there was an earthquake or something and some of the boxes fell off the shelf, a robot as it currently stands would like a Roomba like drive down that lane and then get stuck and wonder what it's supposed to do In the future. You could train it what happens. If that happens, it can do different routes. But, more importantly, you don't actually have to simulate it in real life for it to be trained. You can safely do it in a system, that's, in a digital twin, so that you don't have to accidentally hurt somebody. While you're just trying to prevent people from getting hurt, you can run over your digital human beings and understand what's supposed to be done in that situation. You can drive cars that hit people in a digital simulation and try to train against that, and then you don't have to do any of that in real life, where often cases that's illegal Hopefully, in every case that's illegal. So I found that kind of compelling.
And then, of course, jeff, you and I both waxed ecstatic about the work that's being done in healthcare, particularly when it comes to molecule creation and protein synthesis and all of that kind of medical modeling, but at the same time waxed distraught, I guess, about this idea of talking to a digital nurse who I could say I took St John's Wart this week and I'm about to take some new anxiety medication. Do you think that's a good idea? And boy, I hope that digital AI nurse would say no, no, no, no, no, no, st John's Wart is going to make your body absorb five times the normal amount of that anxiety medication that you would otherwise. So do not take the two. But can we count on that? I don't know if we can count on that. So that part I was not too jazzed about.
1:21:31 - Jeff Jarvis
But no, what's your doctor's name? Dr Jim? No, dr Bill, dr.
1:21:36 - Paris Martineau
Dan and you guys you know terrible news. I received word last week while we were recording the show that Dr Dan is leaving New York. I'm going to have to find a new doctor and I'm really devastated about it.
1:21:48 - Mikah Sargent
That breaks my heart because I feel for you. So much.
1:21:52 - Jeff Jarvis
Where is Dr Dan going? Why would he leave the center of the universe? I?
1:21:55 - Paris Martineau
don't mind to ask him the center of the universe.
1:21:58 - Mikah Sargent
It's my life, it's okay. No, that's fine by me, I don't care where the center of the universe is. That's fine.
1:22:04 - Paris Martineau
The center of the universe is a one medical office in Tribeca.
1:22:08 - Mikah Sargent
I just, if I could afford to, I would. I just I feel it because the doctor I had in Missouri she was in stink incredible, she was amazing and even when I left the city I was like three hours away. I would drive back to go to that doctor because she was so amazing. If I could afford to and also not feel guilty about you know, if I could tailor, swift it and not care about flying in a plane so much, I would go back to her even still as my doctor. But yeah, that's just not. Do you know where he's going?
1:22:40 - Paris Martineau
No, I've got to find out. I haven't. I've been too distraught.
1:22:43 - Mikah Sargent
We'll move Dr Dan Dr. Dan.
1:22:47 - Paris Martineau
I love Dr Dan. I can't leave New York.
1:22:50 - Mikah Sargent
Can't leave the center of the universe. I understand.
1:22:52 - Jeff Jarvis
I can't so he used some interesting phrases too. He talked about the omniverse a lot, dr Dan.
1:22:59 - Mikah Sargent
Yes. The omniverse Always Part of this, dr Dan the universe that consumes both meat and plants. Dr Dan, I guess, so, I guess, so, dr Dan, that's the joke that Jeff made. Dr Dan, that is yes.
1:23:08 - Paris Martineau
Dr Dan, the omniverse, dr Dan. After a certain number of weeks, our brains will just become one Dr.
1:23:15 - Mikah Sargent
Dan, there's a GPT trained on both of your responses in the works right now.
1:23:20 - Jeff Jarvis
Dr Dan, we were replaced. That's fine. He talked about accelerated computing, which sounded a lot like effectively acceleration to me. He talked about when you get your model going, you do this, you do that, you do that, then you guard rail it, dr Dan.
1:23:33 - Mikah Sargent
Yeah, guard rail it.
1:23:34 - Jeff Jarvis
Dr Dan, then you guard rail it. Dr Dan, then you guard rail it. He talked about being not just a chip foundry but an AI foundry making foundry. But then the one that got I think both of us were there and he else. He said about robotics, quote everything that moves will be robotic.
1:23:53 - Mikah Sargent
Dr Dan. Let's just sit with that for a moment, dr Dan. Everything that moves will be robotic, dr.
1:23:57 - Jeff Jarvis
Dan. Well, I don't have to do my exercise anymore. Dr Dan, that sounds like a curse Dr Dan, they did actually show Dr Dan what would put upon you, dr Dan.
1:24:03 - Mikah Sargent
They showed a physical therapist. That was not human Dr Dan.
1:24:08 - Jeff Jarvis
Yes, a robot physical therapist yes, dr.
1:24:09 - Mikah Sargent
Dan, and I just thought that thing is going to make. It's going to be like you think. You can only stretch this far. I will show you, Dr Dan Stop whining Dr Dan.
1:24:19 - Paris Martineau
Stop this crazy thing. Dr Dan, stop whining. Dr Dan, stop whining. Dr Dan, didn't a robot famously give you a medical exam or something? Dr Dan who? Dr Dan?
1:24:26 - Mikah Sargent
Captain was not Leo Dr Dan, I was like.
1:24:28 - Speaker 6
I hope that's not about me, dr Dan. That's not about me.
1:24:32 - Mikah Sargent
Dr Dan, because I thought Leo told some fake story.
1:24:34 - Paris Martineau
Dr Dan.
1:24:34 - Mikah Sargent
I guess it must have been Leo, because Dr Dan Leo has gotten a full body scan from a machine, but it wasn't a robot that he like went in. It's like a full body MRI situation that he's done, dr Dan, but I don't know if he's gotten a robotic body, dr Dan, although I wouldn't put it past him. Frankly, while he was on that trip with the AI person. Maybe there was a robot there, dr.
1:24:55 - Paris Martineau
Dan Anything could have happened, dr Dan.
1:24:57 - Mikah Sargent
Anything could have happened on that day. Yeah, the acceleration that they were talking about. This I actually I've found this fascinating for a while. There's this growing belief, or understanding I guess it's almost becoming an understanding that Moore's law is falling apart, dr Dan, and that we cannot continue to miniaturize and therefore accelerate the computer technology that we have, and that chart that was once just ramping up every 10 years is not something that we're going to be able to continue to sustain, and so there's need to look at different ways of improving upon computing without running up against our ability to shrink the process from 10 to 7 to 5 nanometers, etc. And there is now.
This is where the belief part comes into play, and especially, as you can imagine, with NVIDIA, is that AI could play a heavier role in figuring out how to improve upon performance, even if we can't continue to shrink the process that we have to make chips. And so that is where that accelerate, accelerate, accelerate came into play, talking about using AI to again improve upon that, and interestingly, I didn't remember there being mention of this, jeff, while we were watching, but after that keynote I saw a news story that this boggles my brain they have, nvidia is having a simulated quantum computer so that people can try, so that companies can quantum compute in a simulated environment. But I'm going how can you quantum compute without a quantum computer? How can you simulate a quantum computer without having a quantum computer there?
1:27:12 - Jeff Jarvis
to ship Sin it.
1:27:13 - Mikah Sargent
Yeah, so I didn't understand it, but I was just wondering if you remembered that on stage. Oh, I didn't remember that. Quantum simulation NVIDIA. Let's see what it says here.
1:27:25 - Jeff Jarvis
Keep on saying that things were not animation, they were simulations.
1:27:29 - Mikah Sargent
Yeah, they were trying to make it very clear.
1:27:30 - Paris Martineau
Okay, I will say all of this quantum simulation, digital twins. It sounds like we were in the beginning of like a bio shock spin off.
1:27:39 - Mikah Sargent
I think you may have thought that too, we're marking on the performance overall. So up to a certain point and Jeff really, I think, nailed this because up to a certain point when somebody is making a mistake or making a few mistakes, it's cringey, right, right, but it's almost like I guess it would be what a bell curve-ish sort of thing. Because once it gets up to a certain point of mistakes made and owning up to those mistakes, it starts to fall out of the cringey graph and back down into the like all sort of earnest graph. And that's where he got for us At first. I'm going okay, this guy literally mentioned on stage that he hadn't rehearsed this and that's oh, this is what we get for not rehearsing, and that at first was kind of like oh, but he just sort of leaned into a lot of it and it got back into it.
1:28:35 - Jeff Jarvis
He did jokes where he thought there was going to be a laugh line and it wasn't, and they pedaflooped.
1:28:40 - Mikah Sargent
They did, they pedaflooped. Yes, it was giving Jeb Bush-.
1:28:44 - Jeff Jarvis
All right, I recycled that joke from what we did, but I couldn't help myself.
1:28:49 - Mikah Sargent
That is the, that's the cringiest thing for me the please clap moment from Jeb Bush.
1:28:56 - Paris Martineau
Please. Clap has has become something more beautiful. It has Please clap has has quantified the cringiness it is. You know, it has become a legend in and of itself.
1:29:10 - Mikah Sargent
The generative quantum egen solver, eigan. Thank you, eigan, thank you. Can someone give me context for what that means? Why is it?
1:29:23 - Jeff Jarvis
called Eigan.
1:29:26 - Mikah Sargent
It's a mathematical thing, Eigan. Eigan, you guessed it. Llms to enable a quantum computer to find the ground state energy of a molecule more quickly, as we all need to do. I find myself wondering, as I sit and drink my coffee in the morning what is the ground state energy of the molecule I glance upon?
1:29:52 - Paris Martineau
I'm always asking myself that.
1:29:53 - Mikah Sargent
Yeah, what is the ground state energy of a molecule? If I ever look thoughtful, it's not because I'm zoning out, it's because I'm wondering what the ground state energy of my this morning's molecule is about. Anyway, sorry, let's sort of get back on track. Let's talk about oh, I did want to talk about this A couple of things to do with open AI. Yes, I know, more AI stuff Did wait, sorry, paris, did you? Oh, I thought I heard you breathe like oh, I had an idea, so business insider contemplating the ground state of a molecule.
Oh, good, good, good, good, did you come up with the answer.
1:30:34 - Paris Martineau
Still contemplating. He was breathing heavily during that process?
1:30:38 - Mikah Sargent
Yes, five eigenvalues more energy to do that. Yeah, we can eat a little more energy, it's true, I've got to.
1:30:43 - Paris Martineau
I've got to really oxidate the blood.
1:30:46 - Jeff Jarvis
She was air cooling her processor.
1:30:50 - Mikah Sargent
So open AI? According to our favorite publication, business insider, is reportedly releasing GPT five this coming year, and this is the moment where I wanted to ask Jeff, you talked a lot about where you foresee AI going, and also not just you yourself, but where you know you've read about how people want to see AI going forward. It was the, the paper that you mentioned. You brought it up several times during our stochastic parents Thank you.
About sort of what smaller purpose built sort of situations. I think you'll have to correct me if I'm wrong, but let me go. I'll get through the question first. In a GPT five situation which is supposed to be quote materially better and also supposed to feature some new well, supposed to have some new features, what do you think GPT five is going to present us? Because it can already do photos, it can already do all sorts of text stuff Through plugins, it can generate music, it I guess it doesn't do video just yet, although we've seen some video. We've seen the preview of that.
I think yeah. So how does a GPT or how does chat GPT's GPT four improved to GPT five, Do you think I?
1:32:16 - Jeff Jarvis
think the thing that would scare the breath out of Paris is if it figured out how to do facts, because it doesn't, and then it can start to.
1:32:26 - Paris Martineau
I think it's kind of a. It's ridiculous to assume that it would be able to, because that would imply a level of understanding and a concept right truth that does not exist with a does not large, but can it?
1:32:38 - Jeff Jarvis
can it call on something and cite it at that level? You know, can it, can it. Can it then mimic the value of a search engine? Let's say, the mimic the?
1:32:47 - Paris Martineau
value. Yeah, it kind of similar to notebook LM.
1:32:51 - Jeff Jarvis
Yeah, treat, treat a defined corpus as something to go to. Yeah, I got that is something they've got to tackle. But I think Paris is completely right. Just fundamentally, these things do not do meaning. There is no meaning to them, thus no fact. I don't know. Paris, what do you think the next generation is of these things, on these super big?
1:33:12 - Paris Martineau
It's probably going to be similar to what we have now, but I believe what was the word it said? Materially better. I bet it'll be marginally better. I bet it will have a more up to date corpus. I bet it will have ironed out a number of common kinks. I bet it'll, you know, be GPT for, but updated after however long it's been, you know, based on all the testing and training they've done.
1:33:45 - Jeff Jarvis
I mean, I think it's like it's it's, it's like a new fault. At some point it gets to be like a new phone. I can't remember whether I have the six or the seven sometimes does it matter? Well, okay, I used to care 3.5.
1:34:01 - Mikah Sargent
Are you asking if there's a difference between 3.5 and?
1:34:03 - Jeff Jarvis
there's differences, but I think at some point it's their differences that make I mean again, just like in the early days of twit, they talked about oh, the new monitor has this many things. And oh God, no, wow, yeah, that's three plugs, right. And everybody got excited about it. And knowing that we don't care the same thing around operating systems, right. We used to talk about oh my God, there's a new version.
1:34:24 - Mikah Sargent
I see what you're saying. It's also at some point you don't care, it's similar.
1:34:29 - Paris Martineau
And I will eat my words if I end up being wrong. Okay, but I think that the I will admit it on the show because I'm sure we'll talk about it when GPT five comes out. But I think that we have been at a point where this technology has improved dramatically. It's been very impressive, and we're going to be at a plateau for a little bit after that until we hit the next big inflection point. I don't necessarily think that GBT five is going to be that.
1:34:54 - Mikah Sargent
Yeah, so you're saying we're at that okay plateau. You think it's going to?
1:34:58 - Paris Martineau
keep getting better and it's still a transformative technology. It's not going to, probably will not increase exponentially. You know in with this release, like I think that if it was an exponential increase in terms of the quality of its output and the things it can do, the sources talking to business insider wouldn't describe it as materially better.
1:35:23 - Mikah Sargent
That's true, it would be like this is incredible.
1:35:25 - Paris Martineau
Yeah, the most insane thing I've ever seen.
1:35:28 - Mikah Sargent
Yeah, that's a good point, I feel, like for open AI or for Sam Altman. The only thing that's an exponential increase for him is AGI. And do you think I'm wrong in that? I'm not saying that it's possible, I'm just saying I don't think he sees anything other than that. The way that he talked, he seems to talk about those improvements. He's very, very AGI, agi.
1:35:50 - Jeff Jarvis
It's part of it. It's literally part of the religion. I'm sorry, I'm going to say it. Get your drinks ready, people. It's test real drink. It's, it's, it's their belief that they are the superhumans making the superhuman machine. It's all ridiculous. It's BS, it's BSD, it's it's means. But they're going to keep saying that and then they're going to set their own standard for what it is they're going to get to a point and say see, we have it now.
And Gary Marcus tried to get a musk to agree to a long bet. The musk said we're going to have AGI by 2029. And Gary Marcus said, okay, then let's do a bet, but that's required standards to determine who wins. So Marcus came up with a bunch of things we've talked about in the show before. Like it can, it can read and understand a novel and they can do other and do error free code and other things.
But humans do write free code I know Well that's part of the question is that they think that the goal is to match humanity. Why would you make a machine limited, yeah, to our weird abilities?
1:36:57 - Mikah Sargent
Yeah, yeah, we talked about this on on Tuesday.
1:36:59 - Paris Martineau
I want my AI to have generalized anxiety disorder, ideally, and depression.
1:37:06 - Mikah Sargent
So I'm looking for I needed to have ADHD like me, like what I think. We want it to be hubid and also to need just that we need to eat to stay alive. I need my AI to have to use the restroom every once in a while because it's got renal system.
1:37:23 - Paris Martineau
Finally an.
1:37:24 - Mikah Sargent
AI that can poop.
1:37:27 - Jeff Jarvis
AI poof.
1:37:28 - Mikah Sargent
Yeah, I mean, that is it does get. Maybe that's what it's happening. It's not that it's hallucinating, it's that it's pooping. We've gotten it wrong this whole time.
1:37:40 - Jeff Jarvis
Well, no, we chat. Gpt five is going to have diarrhea.
1:37:44 - Paris Martineau
That's true.
1:37:45 - Mikah Sargent
It's like, finally, now we've got this critical function on the way to intelligence have you ever noticed that, though in a game and in a movie and in a television show, most of the time no one's ever taking restroom breaks, and I just find that unrealistic. So maybe I do, maybe I'm gonna. You know what chat GPT lets you have custom rules and maybe I will add to the custom rules every once in a while. I need you to take a bathroom break for me and just let me know that you're taking care of business Really good bit is everyone.
1:38:18 - Paris Martineau
So all you submit a prompt. It's like sorry.
1:38:21 - Mikah Sargent
I'm taking a crap right now.
1:38:22 - Paris Martineau
I gotta come back in a little bit. It's like sorry, it's actually gonna be a bit longer. I started playing the crossword, yeah, exactly.
1:38:32 - Mikah Sargent
Oh, how did you know what I do?
1:38:33 - Jeff Jarvis
Last night, says the chat GPT.
1:38:36 - Mikah Sargent
I feel called out. Let's, let's, let's move on to what it's continuing to talk about open AI. But I wanted to mention, speaking of crap, there's a lot of spam in the surprise, surprise, surprise in the open AI store.
So stores I don't even know if it's the right word technically, but open AI, of course not too terribly long ago announced feature an option for its platform that when you're using chat GPT, instead of dealing with plugins you could use custom GPT's and then you could publish them. So this is where Leo, if he was doing the story, would excitedly tell you about the GPT that he created. For what language? Emax, yes, emax, thank you you said that was. I know, yeah, I hear about it every week. Oh, lisp, yeah, Lisp.
1:39:34 - Paris Martineau
Lisp also yeah.
1:39:35 - Mikah Sargent
Is the environment. Yeah for for Lisp, okay. So yes, anyway, that's where he would tell you that. So, because I'm here, I get to tell you excitedly about how I trained a custom GPT that I do not share with other people, thank you. It's not published for Dungeons and Dragons. I use it as an when I'm running some. When I'm running a campaign, sometimes the players do things that are just completely ridiculous, or there's maybe a moment where I want to add a little bit more, and so I will have it. You know it knows characters that are in this campaign, it knows certain aspects, and so I could say you know they're going in the store. Can you give me a really deep and rich description of this store and really fun stuff? It makes them cool outputs, but that's a brilliant use for.
AI. Thank you, I think it's pretty cool, and so that's. That's an example of what this does. People can train it behind the scenes. So I, for example, uploaded some PDFs of different aspects of the adventure and then also told it stuff about who my players are, where we were in this campaign, et cetera, et cetera, et cetera. That happens behind the scenes. Then you get that kind of regular chat and you can talk to it and have it interact.
People are able to publish these and make them available to others to use, and Tech Crunch's Kyle Wiggers, who wrote about this, talked about the different aspects of how it is being used in some troubling ways. I'm going to mention one of these aspects. Last, because I think that it kind of falls in a gray area, but I'll talk about something like academic dishonesty. So open AI explicitly prohibits developers from making developers again loose term here, because you don't have to do any coding to train these. I would not consider myself a developer. I was able to train one. It is. It explicitly prohibits academic dishonesty.
Gpt Store has several different tools that help remove those common phrases that the anti-plagiarism and anti-AI generation tools that teachers use are looking for. So it helps to quote unquote humanize your writing after you've had an AI creature writing. There's also impersonation. There is explicit prohibition against impersonation of people or organizations if you don't have their consent or their legal rights. Yet there are many options for Barack Obama, elon Musk and all sorts of other tools. And then there's a jail breaks section as well, which the idea here is that you can, through different prompt engineering, get the AI to provide responses that it would not otherwise give, because it goes against the policy.
Now, according to Tech Crunch's Kyle Wiggers, it was not. They were not very good at actually getting past circumventing open AI's own guardrails or guide rails. But the last one I'll mention is the quote unquote copyright issue, and this is the one where I kind of had a problem with it, because somebody, for example, made a little bot that lets you talk to Eng from the Avatar series. That's not Avatar the Blue People, avatar the Last Airbender the far cooler thing that's going to get emails don't email me. Anyway, it lets you talk to Eng from Avatar the Last Airbender, and again, this is where transformative work versus copyright infringement comes into play. So I didn't care so much about that one, but curious to hear, of course, your thoughts on the rest of it. It sounds like neither of you are exactly surprised. I wasn't surprised to hear about all of this cruft in there. I have not really found. Oh, I failed to mention one, and I've had a personal interaction with it.
A lot of them try to get you to leave the site and go pay for things elsewhere. So there was a sticker tool which was pretty cool. It would generate you type it in and it would create a sticker that they had trained it so that it would always have kind of that white border around it. Is it called the Esposito? It should be the Esposito tool, but it would automatically generate a link for Sticker Mule and so then you would click on it and then go buy the stickers at Sticker Mule. And that's not a terrible. You know that that's fine. But yeah, a lot of people are using this to try to make money by getting you to leave the GPT at the end and go do something with whatever it generates. So, yes, thoughts on the junk and spam and poop I guess that's in the open AI store.
1:44:36 - Paris Martineau
I think one of my main exposure to the open AI store is when we went through it the other month week on the show to delve into the various AI boyfriends and girlfriends which also are in violation of chat GPT's rules on this? It seems like obviously they're not taking a very strong approach to moderation or enforcing their terms of service.
1:45:08 - Jeff Jarvis
Yes, yeah, and the problem here is that you get in trouble when you start dealing with humans and so chat. Gpt made this thing that went out there and people misused it. You know the lawyer asked the stupid question and did wrong things, but that was his fault. Right Now they're starting to redistribute things that are made by humans and humans are going to try to get around rules and do stupid things and do bad things and they have no idea how to deal with this.
1:45:35 - Mikah Sargent
Yeah, they're using a mixture.
1:45:36 - Jeff Jarvis
I don't pay for it. So I can't get into search and I can't see all the stupid stuff. But I'm not surprised it's there.
1:45:41 - Mikah Sargent
They're using a mixture of automatic, as you might imagine automatic filtering and modern Thank you moderation and human moderation and reports from its users. Because what is it? It's so common that a platform is created and then all of the onus has put on the individuals who use that platform to actually do the work of making it what it needs to be. I'm sure that the automated part and the reports from the people who use it are the largest part of what decides what gets to stay and what doesn't. But yeah, again, I don't think I've used really any of them. Well, I'm going to take that back.
There is one tool that I have used from the GPTs that are publicly available not the ones that I've created myself, but they're publicly available that I have used, and that is a tool called Consensus, and Consensus is pretty cool because it is specifically tied to research papers, academic papers, and so there are times where, as is my way, I'm thinking about some strange scientific thing and I get curious about it, and so I will use Consensus to help me find actual peer-reviewed information about a specific topic, and that actually worked quite well. I'm trying to think oh, I was getting some more information on the latest studies on erythritol recently and it was able to point me to the latest research on it, as opposed to the ones that end up getting shared across the web that are old or have a lot of interesting funding behind them. So, yeah, there was one at least.
1:47:37 - Paris Martineau
A brief aside while we have you here. Micah Patrick just posted in the chat a great use for chat. Gpt, which is he wrote, created Dungeons and Dragons, fifth edition NPC based on Jeff Jarvis and it did make a player character. Actually, I will correct. We've got a name Jeff Jarvis. Race, human class, bard. Correct background story. Jeff Jarvis is a charismatic and well-traveled bard who has journeyed across the realms seeking knowledge and inspiration. Born into a family of scholars, jeff inherited a thirst for knowledge and a passion for storytelling.
His travels have taken him to distant lands, ancient runes and forgotten libraries in search of tales to weave into his performance. My question for you, Micah, is what subclass of bard do you think Jeff would be?
1:48:30 - Mikah Sargent
Oh God, my initial thing is college of poor.
1:48:36 - Paris Martineau
I'm not insulting, I think college of lore is accurate and it is a compliment.
1:48:41 - Mikah Sargent
That's absolutely fair, because the guy has so many references that he just pulls out of nowhere. I really appreciate the one aspect of you is that you are kind of like the human version of a bibliography at times, and I don't know You're very chat GPT like in that. I don't know many people personally who are that good about it, and I really like that about you. It's almost like you've got your footnotes.
You've got your mind palace at the ready and you're just like sifting through and you're like, yes, this was by Jensen Copeland Gurp at all Very cool. So yeah, I think college of lore is appropriate, Absolutely. Also, college of lore. Bards are pretty. You should look it up later, Jeff.
1:49:28 - Jeff Jarvis
So is that me with?
1:49:29 - Mikah Sargent
the arrow on the head. Oh, no, that's.
1:49:31 - Jeff Jarvis
Aang Patrick. Oh, that's Aang. Okay, so that's.
1:49:34 - Paris Martineau
Aang, sorry, that's according to what if you? Ask it to.
1:49:39 - Mikah Sargent
Here is your appearance. According to this, it says, jeff Jarvis is a middle aged man with a rugged yet approachable demeanor. He wears a worn traveler's cloak adorned with various patches and trinkets collected from his adventures. His piercing blue eyes and infectious smile draw people to him, making him a natural leader.
1:50:00 - Jeff Jarvis
I have bad teeth.
1:50:00 - Mikah Sargent
Entertaining.
1:50:01 - Paris Martineau
A natural leader Personality. Jeff is known for his quick wit, sharp intellect, insatiable curiosity. He's always eager to learn something new and share his knowledge with others. Despite his style and background. Jeff is equally comfortable in the taverns and markets where he entertains crowds. This captivating stories and enchanting music.
1:50:24 - Mikah Sargent
Jeff, if you have to talk about Gutenberg if you had to have a musical instrument as a because that's what a bard does in this as well what would your musical instrument be? Do you know it, craig Craig?
1:50:36 - Paris Martineau
Craig your voice. Oh, very nice, really getting your money worth for that. Yeah, seriously.
1:50:46 - Mikah Sargent
Yeah, all right, let's see. What else do we have here?
1:50:50 - Jeff Jarvis
Oh, I was trying to explain it to somebody the other day who knows Craig, and they like what you do. What?
1:50:56 - Mikah Sargent
It's probably a difficult thing to understand. There was a Google what Google Keyword post the other day actually yesterday, called the checkup with Google health and this actually is an alternative for, for Paris is losing Dr Dan, perhaps, perhaps it is, although it uses, I believe.
1:51:18 - Paris Martineau
Google employees use one medical as well, so it's.
1:51:23 - Mikah Sargent
it was actually, believe it or not, thanks to Windows Weekly host Paul Therat that I came across this story. So the Windows show was talking about Google, so I thought I'd bring it to the Google show.
1:51:35 - Jeff Jarvis
There are Windows news today too. We could get to those.
1:51:38 - Mikah Sargent
Microsoft. Yes, we do need to talk about that, which we should get to pretty soon, because then we got to go. I wanted to mention that Google is planning on using gen AI Watch out in a few ways, including fine tuning Some of its Gemini model for so, okay, let me let me be clear. The Gemini model of course has some base level understanding of some of what's involved in the medical domain, but what they are doing is they're using sort of special, specific med poem to the large language mind model that is specifically fine tuned for healthcare to further fine tune the Gemini model to help provide kind of people who would be using that tool.
As we know it is in Samsung phones. It very well may likely be coming to Apple's devices in the future. There's reports that Apple is working with Google on bringing Gemini in some way to iPhones and so with that there could be something in the works there that they are improving upon the capabilities, understanding and breaking down X ray images, ct scans, etc. Plus and I think this is the part that's cool and is maybe not as scary or not necessarily scary, but problematic is a personal health LLM that uses your Fitbit and a lot of the information that you get from your Fitbit and your pixel devices to provide some information to you. So if you can imagine if you've got a Fitbit, if you've got a get off your lazy ass Charvus exactly, but it would go.
you know, excuse me, it would go further than that. It would look at your sleep, it would look at your movements, it would look at all of that. And look, I was talking about this on Windows Weekly my Apple Watch, it tracks my. It doesn't track my sleep. I've got, I've got something else for that. My bed tracks my sleep and my Apple Watch tracks my movement throughout the day. Yes, it's fine, shuckle it up, folks. My bed keeps me cool at night and I love it. So what bed do you have? I have the eight sleep, which is a previous sponsor on the network. Full disclosure. And yes, it's it. Actually, I'm a very, very, very warm.
1:54:13 - Paris Martineau
Does it make you cool when you sleep?
1:54:15 - Mikah Sargent
It's, it's yeah. How so it? It passes water through these little. It's basically like liquid cooling. It passes water through this grid that he's a chip, and it sends it into an external device that has an exhaust fan on the back, and so it pulls in cold air from. Okay, this one's easy. It's very quiet, which is a, which is what the coolest part of it is very quiet when you're sleeping.
Even so it starts working before you go to sleep and you don't even. You don't know, it's not even like, like you don't hear it. It's it's very, very quiet. I don't know how they do it it's it's silent. I have to sleep with a CPAP pump. Oh, so you're already in a oh, so you're already wow yeah. Sorry, you're already looked. Listen, I was like you're already.
But no, with that, though, with the CPAP, and you have atrial fibrillation too, right, yeah?
So this might be especially helpful to you, because it can take that's the point that I was trying to get at was I've got all this data that is being tracked and my device only ever tells me, you know, maybe anything about an eighth of it. For example, it will tell me something stupid like hey, this week you walked upstairs less than you did last week. Or hey, your sleeping breath respiration changed over the past couple of weeks, and maybe that information, when paired with a lot more, actually means something. But on its own it's not very. And so the idea that I could have a personal LLM that's kind of looking at all of this, all of it together and making sense of it in some way I think is pretty cool. And it's not going to be a situation where it's like hey, mikey, you know what you should do right now. You should, you should drink water out of the back of the toilet tank, because I read some study that says if you do that, you'll never get sick again. No, you heard a vet come fry your brain.
Geez, that guy just made up out of the blue.
1:56:18 - Paris Martineau
See, there's still creativity in humanity. That's the Dungeon Master's instinct right there, all those hours. They're building towards that.
1:56:28 - Mikah Sargent
Just wait, my dear, dear players, when you are forced with drinking out of the back of a toilet to keep yourselves alive.
1:56:34 - Jeff Jarvis
Anyway, I like this. Go ahead Go ahead.
1:56:38 - Mikah Sargent
No, you weren't finished. Go ahead. Oh no, I was just going to say that I would not use it for those weird things. I would like to hear, if it says I've looked at your trends over the past six months and it does look like, when you are I don't know I'm trying to think of something that it could actually uh, when you spend more time at home in a given week, uh, instead of going places, then you are more likely to uh, not eat as much, maybe, and your sleep is poorer, or something.
1:57:19 - Jeff Jarvis
I think this effort to turn us into a data set, um to find anomalies, fine, but I don't think there's a lot of use in it. Um, you know, how did I find my thyroid cancer? I had a lump. How did I find my prostate cancer? Because my blood tests said, oh, that's not good. Um, how did I find my aphib? Because I said, oh, this is weird. Um, I had another condition. This is old. This is what happens when you get old and you talk about your conditions.
Um, this was inevitable on the show. Um, I had a weird thing with a vein here. Okay, so I went to four different doctors and it was all the blind man on the elephant stick, because the pulmonary person wanted to see what pulmonary. The uh enterconeologist wanted to see was enterconeology. The other one, and finally it was this it was a set thing and all the symptoms pointed to that. My point being, at some point, what you want to be able to do but better your doctor does it than you do it's, rather than is that you can you can feed all of the symptoms and all of the questions of something into it and it will suggest things to the doctor. That the doctor being the blind man and the elephant, thinking only about her own specialty, wouldn't have otherwise thought of. That's where I see a use for this kind of technology. Basically, we're tracking you constantly is probably going to just make you nuts.
1:58:33 - Mikah Sargent
I understand that there's actually uh, there was a few research, uh, there were a few studies about specifically that concern, especially when it came to sleep, that people were getting poor or sleep because of all of the sleep tracking they were doing. But, on the flip side, what if I didn't have to wait until I felt a lump in my thyroid to know that something was developed?
1:58:57 - Jeff Jarvis
But probably blood tests are going to tell you that, and the and the person who was going to give you that is now in jail for a long time. What's her name?
1:59:05 - Mikah Sargent
Oh, I was like my doctor's in jail Dr Dan's in jail.
1:59:09 - Paris Martineau
No, not.
1:59:10 - Mikah Sargent
Dr Dan. That's why he's going out of state the penitentiary's in New Jersey. But see, a lot of times I, I, I imagine, jeff, you have a couple of well, not just a couple. You've got several things on your site here One, you're white, two, you're a man, and three and I mean this in the most loving way I possibly can because of your age, you are more likely to have some of these conditions and therefore, they're more likely to believe that you would have some of these conditions.
I have a. I have a family member who is black, who is young, who is a woman and who has such a, who's also a very self-spoken person and who, because she is a mother, does not have a whole lot of time to do a lot of her own research. That would require that would be required for her to be able to advocate for herself to the level that is sometimes needed to even get the blood test in the first place, and that is where I see this On the. In the same scope, my partner has had a hell of a time working through the, you know, at different times trying to get tests that ended up being necessary, and you know I'm on the flip side of that in some ways.
Obviously, I've got a few things to play against me, but boy howdy, you do not want to come up against me when it comes to making sure I've got all of my receipts and research done, because I do and, frankly, like, if there's one bit of padding on my back, I'll give myself it. So I'm very good at advocating for myself in a health environment, and so my doctor and I have a good relationship where I can literally message and say I would like to get this test done, hear the reasons why, and it's easy to do, but I feel for the people who you know don't have that, and I think that that's what you raised this yesterday, micah, the other day with NVIDIA.
2:01:03 - Jeff Jarvis
The problem, as you said the other day, is the data. For you, as a black man, that would feed these machines is inadequate.
2:01:11 - Mikah Sargent
Yes.
2:01:11 - Jeff Jarvis
And we know that, yeah, and that's got to get fixed in this process before you're going to have enough faith in it.
2:01:20 - Paris Martineau
Exactly, Right Before it becomes a larger issue, which I mean I agree on both the points that I think having this data can be really helpful for people who are not in a position advocate for themselves or be heard by people in the medical establishment.
But I think also I mean last week on the show we talked about how data from internet-connected cars is ending up in the hands of LexisNexis, which is resulting in higher car insurance premiums for people who didn't opt in for that I fear that if our medical data ends up connected all in once to former own large language models related to our own health, that information will somehow get back to insurance companies and ultimately result in I don't know nightmare black mirror situations where it's like, oh, you've been having bad nights of sleep for the last couple of months, your health insurance is going up.
2:02:19 - Jeff Jarvis
Which comes down to not technology but law.
2:02:22 - Paris Martineau
Yeah.
2:02:23 - Jeff Jarvis
But again, as I said last week, the problem is that government will want to maintain that data for themselves, because they're not our friend when it comes to privacy. And so whom do we trust to create the guardrails against the human behavior of these companies?
2:02:39 - Mikah Sargent
I don't know who I have no answers Paris. I trust no one. Yeah, trust no one, dr Dan. That's Paris' dialism.
2:02:47 - Paris Martineau
Dr Dan could be in charge of medical establishment. Hopefully power won't go to his head.
2:02:51 - Mikah Sargent
All right, we will pause to breathe in and breathe out and we are back with this week in Google or some such, the show where we talk about some such and including, apparently, ai's taken big old dumps. Thank you for joining us here on the show, jeff Jarvis, paris, mark, no, and yours truly, micah.
2:03:19 - Jeff Jarvis
Sergeant redefined the digital dump.
2:03:23 - Mikah Sargent
Oh Lord, have mercy, what else?
2:03:27 - Jeff Jarvis
Leo goes away. And what do you do? You make the show about yeah, scatological information, that's.
2:03:34 - Mikah Sargent
That's what we're going, jeff, let's pull from your section.
2:03:38 - Jeff Jarvis
Oh, okay, I think the the inflection story is really interesting.
2:03:44 - Mikah Sargent
Yes, that's right, yes, so tell us about the inflection and Microsoft story.
2:03:48 - Jeff Jarvis
So story, inflection is a company that was backed by a bunch of people, including Microsoft and Reid Hoffman, and it started the chat called pie, which is supposed to be more human and more empathetic and more interesting. And what's his name? He won't forget the name of inflection. And then today came out an announcement where, when I saw, was inflection announcing they had a new CEO because the CEO was going to Microsoft and as you look at all the stories, what turned out to be is that is that, and you can read about this at the information or you can ask your favorite chat, gpt, to seal it for you.
2:04:31 - Mikah Sargent
Oh snap.
2:04:34 - Jeff Jarvis
Them lines are drawn. So what happened is that that Microsoft, is you remember, wanted to have a new a to I division and it was going to be Sam Altman in the one hour when he wasn't the CEO of open AI, and then he was back at open AI and we didn't hear anything about that. But Microsoft was figuring out it needed an open AI division and AI division to make sure they hedge their bets with open AI. So they managed to hire much of the staff, almost all the staff of inflection, starting with Suleyman, so they now have ahead of Microsoft AI. What happens to inflection, you might ask? Well, it pivots to something else with a smaller staff, with a new CEO who used to be the head of R&D at Mozilla and read Hoffman wrote about this, tweeted about this and he was obviously the puppeteer behind this deal and the deals really kind of amazing, because the read at Greylock and Microsoft and the investors in inflection are being made whole, apparently by licensing deals to Microsoft rather than a purchase.
It appears from what the information tells me. But here I am stealing the information and sharing it with you for free, because that's the kind of bastard I am. So the information informs me. Actually I'm plugging the information right now and doing exactly what Google does. I'm saying go buy an information subscription, it's where I found the best information. So they're going to make the investors whole. But the investors also keep their same equity in inflection. If inflection grows or sells or does other things, they will benefit again. Meanwhile, microsoft gets the best possible people they can for Microsoft AI and Suleyman gets a great job. It's a pretty weird, amazing deal that occurred here, so it just kind of came out very quietly and I found it fascinating.
2:06:25 - Mikah Sargent
I'll be honest, I did kind of lose the thread there between the veiled attacks. I mean.
2:06:36 - Paris Martineau
It seems like what has happened is Microsoft hollowed out inflection because it got most of their senior staff. At the same time, microsoft is doing a licensing deal here at Alprudian District. At the same time, it revealed the staff departures. Inflection disclosed a licensing deal with Microsoft to make its models available for sale on Microsoft Azure and said other cloud platforms would follow. Inflection is also told investors they will fully recoup their investment and more as a result of the licensing agreement. It's unclear what the size of the payout will be and whether it will come in a lump sum or over time. In addition to the payout, investors will keep their equity in inflection, ensuring that they can profit if inflection is bought or goes public, which is a really interesting setup where it's like Microsoft gets all the good stuff of inflection and investors recoup their investment but still hold equity in the skeleton of inflection in case something comes of it.
2:07:48 - Jeff Jarvis
It's a new CEO people respect and then it pivots. I think it basically becomes that job title I was making fun of before. I think it becomes the L-MOPs, whatever, when it becomes a kind of product integrator person thing for LLMs.
2:08:14 - Paris Martineau
Jeff, how do you feel about the kind of consolidation we're seeing in the AI industry?
2:08:21 - Jeff Jarvis
That comes back to the kind of clutch moment I had watching Nvidia, where if big remains everything in AI, then that kind of you have to have the most data, the most power, the best brains, the most money. Then AI becomes the province once again. If you're talking, I think, before we're Paris, with respect, your reflex to trouble was let's regulate, not necessarily my first response.
2:08:53 - Paris Martineau
Enemy of regulation. Jeff Jarvis would have some other thoughts.
2:08:56 - Jeff Jarvis
I want smart regulation, I just don't want dumb regulation.
2:08:59 - Paris Martineau
I want smart regulation that incorporates the opinions of the companies being regulated.
2:09:03 - Jeff Jarvis
Well, that's an issue, and so what's going to happen here, I think, is that the lesson we should learn in any case is if we let things get consolidated in a few bigs, we're a problem. Now, this is where I think the fact that Meta started the track going of open sourcing, lama and Gemma is Gemma and I open source, and Europe didn't forbid open sourcing, which would have been really awful. So the question is how can we and there's and there's what's, the French one, mistral, which is, which is fairly open source how can we put resources behind all of that? And if the administration believes they're going to throw money behind these, they just put $8 billion into an Intel chip fab. Maybe they should be investing in smaller, competitive, open source AI work in universities and institutions, which again is happening in Norway. Why can't we be Norway?
2:09:59 - Paris Martineau
That's why I think this is my colleagues at the information posted a story, I guess, while right as we jumped on to record this that cohere, the open AI competitor. Apparently they were generating about $13 million in annualized revenue at the end of last year, which means that they, in one month at the end of last year, generated only about a million dollars in revenue, which seems very small to me. It seems like many of the companies in the sector are still incredibly immature.
2:10:38 - Jeff Jarvis
Yeah, I think you're right and I still don't understand how, how, opening I guess, the revenue it gets, because how it's really used it's not really really really useful as a model. I don't think it's a toy. So, unless we talk, this is where Leo has to set me straight.
2:10:57 - Paris Martineau
Leo's like Lisp. I honestly you pay for a chat to be tea to run your dungeon master.
2:11:11 - Mikah Sargent
I do actually. So I paid for it in the first place because I wanted to. You know, whenever it first started adding new stuff, we were talking about it a lot. I wanted to show the new features that it has. Now it's kind of mostly familiar with everybody, but I do get value out of it outside of just the dungeons and dragons stuff and especially when it comes to creating new automations and things that I do.
2:11:38 - Jeff Jarvis
Wait, do you? Do you disclose this to your fellow players?
2:11:41 - Mikah Sargent
Yes, yes, yes, yes, yes, that like would you consider that to?
2:11:46 - Jeff Jarvis
be an ethical requirement.
2:11:50 - Mikah Sargent
I don't know. I think if I was playing online in front of like an audience, that would be necessary probably. I don't know.
2:12:00 - Paris Martineau
Yeah, I think if you were playing in front of an audience, maybe, but I also think, like a lot of people use modules for D&D, which is for those who I guess don't know, modules are kind of like a pre-written campaign structure. It's a book that could have everything you want to know about a particular world or an adventure and do you kind of base your game around that. And this is similar. I mean, what is it drawing from your chat?
2:12:28 - Mikah Sargent
Mine actually is drawing partially from modules as well as information about them. You pay other people for them. Yeah, you have to buy them. It's from. Okay, wizards of the. Coast.
2:12:39 - Paris Martineau
It's from Wizards of the Coast specifically. Don't turn off.
2:12:43 - Mikah Sargent
You have to.
Well, yeah, that's where the adventure came from, so that is owned by Wizards of the Coast or that's where I've purchased that from has been, and then parts of that also have come from individual creators.
You know when I've added different things to the story, and then of course there's stuff that I've added myself and have. That's the hardest part, because if it's something that I've created myself, like this is maybe a long, it goes into the weeds. But essentially, there was a part of the campaign where my players needed to learn how to use griffins to fly from one place to another, and so on the spot, I just came up with this whole set of instructions that this sort of grouchy person was teaching them how to do, and so I'm writing it down in the moment as I'm inventing what the rules are for learning how to fly and train griffins, and I needed that to end up into it. So having to kind of take all of that and summarize it and put it together and make it so that the thing new is a lot harder than just taking the PDF that you've purchased you know, the two-page PDF about a specific location that you purchased and just uploading that into the system.
So that's so interesting yeah when it comes to, yeah, I think, if I ever so, I do play in a campaign that is made available to the public. It's part of the incomparable network, which is a friend of the network, jason Snell's network, and he has a podcast series called Total Party Kill, and there are many different adventures and I think that if I was the dungeon master on there and I was using chat GPT, that probably would be something I would disclose in that case, just again, because it's public facing. But yeah, I don't know that it's necessary in private per se.
2:14:48 - Paris Martineau
I mean, I think also I don't even know that it would be because I listen to a lot of popular like actual play podcasts and whatnot, for instance, when I listen to his Not Another D&D podcast and often the DM on that will go into kind of their aftershow. Oh yeah, I kind of based this monster off of something from Kobold Press, which is a third party.
2:15:08 - Mikah Sargent
Yeah, I mean all of creation is kind of curbing off of other things. You're absolutely right, and there's no set of rules where, at the very beginning of any new thing, we go. I purchased this from it's almost it's a little bit like robotstxt. So I worked to raise money for St Jude Children's Research Hospital with the Dungeons Dragons campaign and as part of that it was a live stream where people they would make donations and, depending on the size of their donation, they could impact the game in a certain way. Oh, that's so fun.
So it was a lot of fun because people would pay a amount of money to bring a monster into the campaign or bring an adventure to join my players. And I chose to, but was not required to start out by saying by the way, this epic level adventure is from this person, you can buy it here, and I've adapted it as such. And then I also did one that wasn't a money raising venture. It was how the what's it called how the Lich stole Christmas. But I didn't want it to be Christmas because A I had one player who was a, who was Jewish and was a practicing Jew, and I did not want them to feel like out, and so I completely changed the name how the Lich Stole Murthmus.
But I also made it clear this comes from this adventure and we took it and adapted it, and there's a lot of adaptation that goes on and I think there's a lot of just general. It's all good and really it's only ever been since Hasbro came into it that things have gotten particularly messy and you may have seen I don't know, jeff, if you have any attention in this space, but you may have seen, and maybe Paris you did see when Hasbro did try to change the tables.
2:17:05 - Paris Martineau
When they tried to change the creative kind of their version of Creative Commons licensing.
Specifically what is interesting here is, like D&D and correct me if I'm wrong, because I'm just kind of riffing off what I remember from ever this was D&D has always had some sort of licensing structure.
The thing I had mentioned earlier about like a DM of a popular show saying, oh, I got this monster from something called Cobalt Press.
That's a third-party company that exists to make like, let's say, monsters or NPCs or campaign modules that fit into the world of Dungeons Dragons, a system owned by a company called Wizards the Coast which is now owned by Hasbro. And they're able to do that because of a specific kind of licensing structure that says like, hey, you can create content in our universe using our rules and whatnot. That's totally fine, but they tried. The community was in an uproar sometime last year, I believe, because they tried to change it, essentially to get larger companies that have popped up where their whole business model now is creating content in these worlds using these rules, and saying, if you make over, I think, some X amount of millions of dollars, you have to pay licensing fees, and people were up in arms. I mean, it's kind of a similar issue in regards to copyright that we talk about in this show all the time with tech products and the thing is very quickly.
2:18:28 - Mikah Sargent
It becomes, I would argue, a fair use. What is the exact term? It's an alteration, right, it's? Well, here is an alteration. What is the verb Transformative?
2:18:41 - Jeff Jarvis
Transformative, transformative use, yeah it very quickly becomes transformative.
2:18:45 - Mikah Sargent
Use the moment that any of your players does anything, because that is not at all written into what the adventure provides, and so and you have to react to that, so they become a co-creator then yes, yeah, which is something we don't account for enough in a world where we can collaborate far more than we ever could before.
2:19:10 - Jeff Jarvis
Yeah.
2:19:11 - Paris Martineau
And something I think that is interesting, because someone in the chat while we were talking about this posted like oh, I'm surprised D&D is still even a thing anymore. D&d is a huge business, one of, I think, as Patrick rightfully pointed out, the top creator on Twitter for the last many years is this company critical role that makes tens of millions of dollars every year just by running live stream D&D campaigns on a weekly basis.
2:19:39 - Jeff Jarvis
Twitter or Twitch. On Twitch If Leo was making millions of dollars to that he'd be fine. He'd stop pushing for the club.
2:19:49 - Mikah Sargent
Speaking of, let's take a quick break. Twittertv slash club Twitter $7 a month, $84 a year. Consider joining the club. When you do, you help support what we do here on the network. You help us keep talking about AI doing weird things and Dungeons Dragons and watch Jax Javros and Paris Smart. No, fight each other over copyright and fair use no fight.
No, not you. Jax Javros fights. It is discourse. So join the club. You'll gain access to a bunch of awesome benefits. Twittertv slash club Twitter has all of the information for you. I won't go into it right now, but there's lots of great stuff, including every single show ad free. It's just the content $7 a month, $84 a year at Twittertv slash club Twitter. I think we're nearing the end of this episode of twig, so I want to ask both of you if there is any other story that you absolutely want to make sure we get to before we go to the picks of the week. If not, I'm on picks by random. Yeah, let's go similar you go.
2:20:59 - Paris Martineau
All right Listeners of the show may remember some months ago in which I introduced Jeff Leo and others to the phrase to the word Riz, by showing them the video I don't know if you've seen like Livy just rizzed up baby Gronk and from the mind that brought you that video, we've got an important update on line 83, which is baby Gronk just got revenge on Liz Livy.
2:21:29 - Mikah Sargent
What I don't know what any of this means, is this more crypto?
2:21:34 - Paris Martineau
No, strangely it is not. This is sports. We have to play the video from the beginning, just the first, I don't know like. I think it's 20 seconds with sound, ideally.
2:21:49 - Mikah Sargent
So well here we go.
2:21:49 - Speaker 4
Here we go. Baby Gronk just got revenge on Livy. Livy has been cheating on baby Gronk with Paul Schien. She's even been watching his baseball Confusion is the right way. Is baby Gronk fair?
2:22:00 - Mikah Sargent
enough.
2:22:02 - Speaker 4
Sidney Smith. To make Livy jealous, baby Gronk took Sidney out to dinner. He even hugged her. Is baby Gronk?
2:22:08 - Mikah Sargent
He even hugged her he even hugged her.
2:22:11 - Paris Martineau
He could be the Rizzler of the year, guys.
2:22:14 - Mikah Sargent
He hugged her. He could be the Rizzler of the year. Jeff, Did you hear that? The Rizzler of the year? Of the year.
2:22:20 - Paris Martineau
The point of this. There is no point. It's simply that this man says words like they mean something and they don't.
2:22:29 - Mikah Sargent
But you have to respect that a little bit, don't you? When you say something with enough gloomption, you just believe that it's real and you just Movie and you really do, and you're like did baby Gronk get revenge on Livy?
2:22:46 - Paris Martineau
Is he the next Rizzler? He might be the Rizzler of the year even.
2:22:50 - Mikah Sargent
Oh wow, it's a great question. So can I do? There is clarification to be done here. The most important clarification that I have is is baby Gronk old enough to be dating Lizzo? No, what's her name? I don't think there's any dating going on All correctly.
2:23:09 - Paris Martineau
the first video he said with the same disturbingly strange affect. Baby Gronk just rived up Livy, and baby Gronk is a child football star well known on TikTok. Livy is maybe a college cheerleader of some sort.
What really happened is baby Gronk and Livy met and had a brief side hug that was captured on some sort of video. But this man's job is to kind of do sports entertainment news about things that clearly don't matter, so he created this incredible story. Well, I mean, I think part of it is just the absolute absurdity of a video that begins earnestly with baby Gronk just Rizzed up Livy In a similar sense to this what's arguably even more absurd is that I genuinely thought it was a cryptocurrency moment.
2:24:12 - Mikah Sargent
I think that speaks to the level of absurdity that tech embodies.
2:24:19 - Jeff Jarvis
I like that we began and end the show with Paris's linguistic explorations.
2:24:25 - Mikah Sargent
That is what it is You're right.
2:24:26 - Paris Martineau
These are the things that haunt me. Yes, this specific section of my Twitter bookmarks that are just things that haunt me that I could talk about in this show. I appreciate you guys for joining me on this.
2:24:37 - Mikah Sargent
The things that will haunt us as well, you know, I think, oh, why don't you talk about what you're on there right now, mr Jarvis? Some booing, and we'll end on that, but I'm 74.
2:24:53 - Jeff Jarvis
Yeah, this was a cultural moment At South by Southwest, which itself is obnoxious. They think that California is more obnoxious because they booed any references to technology in the sizzle reel that South by Southwest put up. It's not great video, but you'll hear that at any mention of technology or AI, the Texas crowd, the South by Southwest techie formerly the techie crowd now cooler than techies booed Silicon Valley.
2:25:35 - Paris Martineau
This is very funny, coming from a conference that is, like, funded by the biggest defense contractors.
2:25:41 - Jeff Jarvis
Exactly.
2:25:50 - Mikah Sargent
Oh my God, yeah, that's pretty great.
2:25:56 - Jeff Jarvis
I love that it.
2:25:57 - Mikah Sargent
Yeah, boo Boo. That would have been troublesome during the Nvidia conference. She's Louise. All right, believe it or not, we're going to do one more breath before we head to the picks at the end of the show. Air, cool your chips. Air, exactly, all right, it's time for the picks, stuff and things of the week. Paris Martina, what is your thing of the week?
2:26:29 - Paris Martineau
My thing of the week. I don't. You're the only person who have any undripped possible possibility of understanding this, micah. Have you watched anime at all and do you know what I'm saying when I say neon Genesis Evangelium?
2:26:41 - Mikah Sargent
So I will say that I have watched some anime and I have several friends who really like Evangelion. I do not know what the neon Genesis part of that means, so Evangelion is a shorthand for neon Genesis Evangelion.
2:26:56 - Paris Martineau
I'm learning.
But, yes, the same thing. I have not watched Jantime since I was a child. I used to be really into it. Had no real interest in getting back started.
This might be my only foray, really, but I've had some friends mentioned this series to me time and time again. It's a series that originally aired in the 90s or something like that. They describe it as the best anime of all time, if not one of the best television series of all time, and I had kind of heard that and rolled my eyes a bit, but I was like, oh, whatever, the episodes are like 20 minutes, it's on Netflix, I'll give it a watch. It is genuinely phenomenal. It's a deconstruction of the. This trailer I included is not going to make sense, but it's beautiful. It was moving at the deconstruction of the mecha genre, even if you have no idea what's up with anime and how they recommend it. And it ended up becoming very timely because the series is maybe about like 20 some episodes. It is no known for having one of the strangest, like most artistic post-post-modern absurdist endings in anime ever. It's only 20 episodes.
Yeah, it's like 26 or something.
2:28:10 - Mikah Sargent
I can watch that all.
2:28:12 - Paris Martineau
You can definitely watch it. All the episodes are like 25 minutes long.
2:28:16 - Mikah Sargent
The way people talk about it. It seems like it's this huge epic story.
2:28:19 - Paris Martineau
That's what I thought as well, but it is. It's a real treat and it ended up being a great time for me to watch it, because just this week I finished this series on Saturday by accident, and it turns out on Sunday, and today, on Wednesday, the studio was releasing. There was like a movie made as kind of a coda to the series and it was being shown. It's maybe been like quite at least a decade or two since the movie first came out, but they re-released it in American theaters for the first time ever on a Sunday and Wednesday. So I went to the theaters in Manhattan and saw the end of Evangelion and it was a delight.
2:29:05 - Mikah Sargent
Nice, I'm told.
2:29:08 - Benito Gonzalez
Yeah, so just that show is actually. You know it's up there with, like Miyazaki. It's kind of transcended anime already. It's like its own.
2:29:14 - Paris Martineau
Yeah, yeah, I would hesitate to call it anime. I mean the first maybe like five episodes of it seem. It's more just kind of like getting used to the concept. At first I was like this seems like a pretty normal anime. You got like a kid inside a giant robot. End of the world. Whatever it goes places, I'd put it up there with some of my favorite absurdist, nihilist 1970s films.
2:29:37 - Jeff Jarvis
Well, is that kind of blurb Absurdist nihilism?
2:29:43 - Paris Martineau
I was not expecting it to be as interesting artistically as it was, but Well, I'll have to watch it then, knowing that it's on Netflix, check it out, I'm not a huge, huge commitment, because, yeah, that's the thing is, I was some of them.
2:29:58 - Mikah Sargent
I don't have the time to get so deep into all of the lore.
2:30:01 - Paris Martineau
Like Naruto, that's hundreds of episodes. Yeah, I don't have time for that.
2:30:05 - Mikah Sargent
Ain't nobody got time for it Exactly, but we do have time for Jeff's stuff.
2:30:12 - Jeff Jarvis
So you all make fun of me, but I'll have. You know I'm a New York Times trend story that people are not going to get rid of their phones, even though I didn't know.
2:30:25 - Mikah Sargent
At&t is talking about eliminating them in California eliminating and after the landlines the rain lands come to your house and they're going to take them away from you.
2:30:37 - Jeff Jarvis
So, and your guns?
2:30:39 - Mikah Sargent
Okay, don't say that that's good.
2:30:43 - Paris Martineau
You've heard it here first, guys.
2:30:45 - Mikah Sargent
Jeff Jarvis says they're coming for your control. It's all going.
2:30:48 - Jeff Jarvis
Your landline phone and your pills. So they came for the landlines and I did nothing, so I didn't notice the staff in the story. 73% of American adults lived in a household in 2022 with no landline but at least one cell phone. That actually surprised me. It's gone that far. My, my, my cohort has died off. Nearly 90% of Americans aged 25 to 29 reported that they used only cell phones, compared to less than half of Americans over 65. So they're taking away grandma's phone.
2:31:26 - Paris Martineau
Wait, do you actually make calls on that landline?
2:31:30 - Jeff Jarvis
No, here's the weird thing about it.
2:31:32 - Paris Martineau
Okay, here you are talking this big game. When is the last time you picked up a call from that phone, jeff Well pick up never because.
2:31:39 - Jeff Jarvis
well, when our son who lives in New Hampshire calls, he calls the landline, because we can put that on speaker and my wife's phone might be here.
I might be there. This thing, paris. There's this amazing thing here. You won't believe this. You'll believe it. Blow your mind. They call them extension phones. The same phone number rings in multiple rooms. It's quite amazing, wow. And so we can pick one up wherever we are and we can talk to our son, right, so that's fine. So we use it for that. No, all the calls are spam calls, every single call, and my wife will yell at me. She said don't answer that one. And I said no, no, it's gonna be a real good. No, it's always. My trick with them is I always pick up the phone and then I say nothing for 10 seconds.
2:32:26 - Paris Martineau
See that doesn't do it because they're noting that you answered the phone and that it's an act of lying.
2:32:30 - Jeff Jarvis
It's like an old person. They're gonna call you again.
2:32:32 - Paris Martineau
Yeah, they're gonna convince me that grandchild is an Niagara Falls you should pick it up and then give that person on the other line all those gift cards they've been asking for. It's really necessary.
2:32:43 - Jeff Jarvis
Well, the other thing I do is I then, if I do say something, I say something with a very gruff Jeff voice, hello. And then sometimes it's actually somebody who's wanting to call and I sound like I'm a rude asshole. So that's the other problem with it. No, we're on good terms.
2:32:58 - Paris Martineau
So this is your case for landlines.
2:33:01 - Jeff Jarvis
It's just we just got an inertia man. Inertia we had until not many years ago. We had two lines, one full time for the fax machine. Wow.
2:33:10 - Benito Gonzalez
I think landlines are just being propped up by the home alarm industry at this point, because they all still require landlines.
2:33:17 - Jeff Jarvis
Plus. No, they don't all. We have somebody out and they don't use that anymore. They use the ADT, google use the cell.
2:33:24 - Mikah Sargent
I think a large part of it is just regulation, because I remember just the other day on not the other day, but a couple of episodes of Ask the Tech Guys ago someone wrote in and was talking about how AT&T was holding a conference about trying to get rid of landlines. But it was required by the FCC that a corporation in a place where there's no other way of calling 911, if you're the only body that can be used to call 911, in order to leave that area, you have to hold a series of public conferences and whatnot to say we're leaving, we want to leave, and then it also has to be agreed upon that they can leave, otherwise you have to stay. So I know that that's a big part of it is overcoming that regulatory hurdle. It seems from the FCC that is saying look, if you're sorry, but you're the only one here, is giving these people the ability to call emergency services. So you kind of got to stay here.
2:34:33 - Jeff Jarvis
But you are keeping it alive yourself. This is not a copper line anymore. Obviously, this is called our fiber. At&t said the number of copper landlines known as POTS fell by 89% from 2000 to 2021. Quote like blockbuster rentals and Kodak film, pots has fallen from technological primacy to effective obsolescence in the course of a generation. At&t wrote an application to the California PUC. If my 97 year old departed father, we got him a cell phone but, oh my God, he couldn't hear on it. He couldn't work out. We had to get one of those old folks phones that did the text for you and all that stuff.
2:35:17 - Mikah Sargent
Oh yeah, was that TT something, something? No, it's kind of like that, it's a little different. Got it.
2:35:23 - Jeff Jarvis
In Florida. There's so many old people in Florida. They give them away for free.
2:35:26 - Mikah Sargent
Wow. Wow. That's where all the Florida tax dollars are going, not to gator clean up or wait. They have crocs there.
2:35:35 - Paris Martineau
I let those you know Rome free.
2:35:39 - Mikah Sargent
What else in your stuff pile, Jeff?
2:35:42 - Jeff Jarvis
No, that's enough.
2:35:43 - Mikah Sargent
That's enough. Okay, he's had enough. My pick we've got things, stuff and pick. My pick is a little app that I just came across called Minders, and I thought it was interesting because it was first and foremost designed as a Mac app, with the kind of iPhone app secondarily, and it is. It kind of works how my brain works, which is that throughout the day I have thoughts and I look at different sites and I see different products and I do different things and I the way that I kind of keep hold and keep track of all of this is that I just have a text message conversation with myself and so I just send things to that text message conversation, but it's not incredibly searchable, it's not great for kind of scrolling back and seeing different things, and somebody created this little tool that works almost like a private social media. It's just you and what I liked about this idea you finally get the answers you want.
Yes, what I liked about this idea Finally, only good posts, only good posts is that I have almost exclusively, or almost entirely, not exclusively, almost entirely stopped posting anything other than the occasional link to something that I'm doing or what have you.
Because what would happen every time is I would think about posting something and then the five to 10 expected responses that I would get would fly through my head. And so then I'd start rewriting the message so that it had all of the proper hedging in place, and then I'd think about the next set of messages that I would send to make sure that I was covering all of my bases for all of the ways that it could be misinterpreted or changed or what have you, or the questions that might be asked. And by the time I was done with that, I'd say I'm exhausted and I don't want to post it. But I thought what if I could just send those little funny moments I have into this little feed and I could just have them there, and every case I could scroll back and go oh yeah, on that day I was thinking about how there may be good bacteria in the back of a toilet tank. What a bizarre idea. And then I'll think about.
What is it with you and the toilet tank, geez? And then I'll just think about this great conversation as part of that. So, yeah, this is like a little social media for yourself that you can post photos, and it's got like link cards as well, and you can set up different streams. So maybe you have like for Paris, you have the things that haunt. You could go there in one stream and then you could have all of your thoughts about who's going to replace Dr Dan, I guess. And anyway, I thought this was a pretty cool little app called Minders, and what's great about it as well is that it is very privacy-minded. Minders is it's all iCloud syncing. So, yeah, you might check out that app, available for free in the app store.
Folks, I think that brings us to the end of this episode of this Week in Google. We had a lot to talk about today and we only got to about half of it. So no, I kid, it was a great conversation with you both. Thanks for having me here with you. Good job, micah. Thank you so much. And Paris, if folks want to keep up with what you're doing and want to follow along, where should they go online? And do you have anything you'd like to plug?
2:39:15 - Paris Martineau
Can follow me on Twitter at Parismartino, on blusky at Parisnyc, or reach out to me on Signal with interesting tips about companies at a martino01.
2:39:32 - Mikah Sargent
And of course Jeff Jarvis. What about you? Where should folks head?
2:39:37 - Jeff Jarvis
GutenbergParenthesiscom, where there are discount codes still for my two books out now, Magazine and GutenbergParenthesis. And thank you very much.
2:39:46 - Mikah Sargent
Excellent. Thank you both. You can find me online at Micah Sargent on many of social media network or the best thing to do is head to chiwawacoffee. That's C-H-I-H-O-A-H-O-Acoffee, where I've got links to the places I'm most active on line, if you would like to. If you're listening to this episode and you're not subscribed, you can subscribe to Twig and also find the show notes, all of those beautiful links over at twittv slash twig. I guess that's a long one in comparison to what I'm used to. Yeah, twig, twittv slash twig. We'll get you to the show page where you can subscribe to audio or video formats. This show records live every Wednesday around about 5 pm Eastern, 2 pm Pacific, 2100 UTC. If I'm hosting the show, if Leo's hosting it, it's sometime after 2 pm Pacific. Don't tell him. I said that. Thank you so much for tuning in. We will see you or they will see you for another episode of Twig next week. Bye-bye. Bye. Starred by.